	Nitroglycerin was one of the first medicines used for the treatment of angina pectoris. Murrell noted in 1879 that administration of the drug would provide quick, if fleeting, relief from chest pain.1 Since then this relief has been used as a tool to predict the likelihood of obstructive coronary disease. The Diamond and Forrester criteria go so far as to include it as one criterion to define typical angina pectoris.2 The mechanism of this relief, however, is multifactorial. Nitroglycerin acts on vascular smooth muscle cells to stimulate an increase in cyclic guanine monophosphate, which results in a dilation of veins, arteries, and coronary arteries. The relief of chest pain, however, is not likely to be the result of an increased blood flow through obstructed coronary arteries. Instead, in the acute setting, nitroglycerin via systemic vasodilation is thought to decrease preload, thus decreasing myocardial work and oxygen demand.

	Henrikson and Chandra-Strobos (page 22) investigated the validity of chest pain relief from nitroglycerin as having predictive or prognostic value in active coronary artery disease (CAD).3 They prospectively evaluated 459 patients who had chest pain and received nitroglycerin. These patients were then classified as being “nitroglycerin-responsive” if they had a greater than 50% reduction in pain (subjective scale) and were simultaneously designated as having “active CAD” or “no active CAD” on the basis of appropriate symptoms with at least one of the following: elevated troponin T, coronary angiography with 70% or greater stenosis, a positive exercise stress test with or without imaging, or a clinical diagnosis made by typical symptoms in the setting of known severe disease. To be classified as no active CAD, patients also had to be free of cardiovascular events in the follow-up period of 4 months. The sensitivity of response to nitroglycerin in evaluating for CAD was 35% while the specificity was 59%. Changing the definition of responsiveness from 50% to other percent reductions or by an absolute reduction on a pain scale did not increase the predictive power of nitroglycerin responsiveness. They also noted that there were no significant outcome differences between the nitroglycerin responsive and unresponsive groups.

	The authors acknowledged that the study was limited by the subjective nature of the pain rating scales. Also, the study did not standardize the workups of the patients once they were admitted to the hospital. The hazard is that those patients who responded to nitroglycerin were then more extensively evaluated for CAD. This investigation was conducted in the acute setting and on patients with acute chest pain; thus, extreme caution must be used in extrapolating these findings to patients with chronic chest pain in an outpatient setting. It would be dangerous, based on these findings, to dismiss an escalation of chest pain that is relieved by nitrates in a patient with known CAD.

	This study highlights the complex nature of evaluating chest pain. Pathology, symptoms, and therapeutics do not follow a simple formula, and the findings clearly support the idea that pain relief from nitroglycerin does not have a significant role in establishing an acute diagnosis. However, it would also be imprudent to totally dismiss the use of nitroglycerin in the acute setting. While a physician may not be comforted by the response (or lack thereof) to nitroglycerin, alleviating a patient’s discomfort without detriment remains a worthwhile goal.

Articles

Cardiology Review Online

Chest pain relief from nitroglycerin: Very little comfort

	A 54-year-old man with a history of hypertension and a family history of coronary artery disease (CAD) presented to the emergency department with intermittent chest pain over the previous 6 hours. He characterized the pain as squeezing, substernal, lasting for 5 to 15 minutes at a time, and not clearly related to exertion. He had no history of CAD, and the only medication he was taking was an angiotensin-converting enzyme inhibitor for hypertension.

	While in the ambulance, the patient had an episode of chest pain, which he rated at an intensity of 8 on a 10-point scale. He was given a sublingual nitroglycerin tablet by the ambulance crew, and 5 minutes later, he rated his chest pain as 1 on the same scale. In the emergency department, his electrocardiogram (ECG) was normal, and the results of his first set of cardiac enzyme tests were negative. Because of his moderate risk profile, the patient was admitted to the telemetry unit on the medical floor, with a provisional diagnosis of “rule out myocardial infarction [MI].” He said he had had a large meal before the onset of the chest pain, but reported that he did suffer from “heartburn,” and his current pain was unlike that of heartburn. Serial enzyme tests and ECGs showed that the patient had not had an MI, and he was discharged the following afternoon with a diagnosis of gastrointestinal chest pain. He underwent a stress test the following week and completed 9 minutes 15 seconds of the Bruce protocol, without any symptoms or electrocardiographic changes.

	The patient had no further episodes of chest pain during the 4 months following his admission. Although he received chest pain relief with nitroglycerin during his initial presentation, he was not diagnosed with active CAD.

Case report:  Chest pain relief  by nitroglycerin

	From the Sección de Cardiología No Invasiva, Servicio de Cardiología, Hospital General Universitario Gregorio Marañón, Madrid, Spain

	In patients with atrial fibrillation, the left atrial appendage has been implicated in the formation of atrial thrombi and, as such, is involved in the origin of about 25% of all strokes.1,2 In previous studies, it has been suggested that greater risk of embolism may be indicated by markers such as the size of the left atrial appendage, reduction of left atrial appendage flow velocities,3,4 and detection of spontaneous echo contrast in either the left atrium or the left atrial appendage.5,6 Based on this decisive role of left atrial appendage, we assessed whether stroke risk correlated with surgical removal of the left atrial appendage.

	We retrospectively studied 242 patients who had previously had mitral valve replacement and who were referred for echocardiography at our institution. Patients with other possible causes of systemic embolism were not included in the study. A total of 205 patients took part in the study, of whom 130 were women (mean age, 62.4 years). Patients with biological valves received anticoagulation or antiplatelet therapy, and patients with mechanical valves were given warfarin (Coumadin). Occurrence of a stroke or embolic transient ischemic attack occurring 48 hours after the surgery were the outcome measures.

	Transesophageal echocardiography (TEE) was used to carefully examine the left atrial appendage and the left atrium for the presence of thrombi and left atrial appendage spontaneous echo contrast. The lack of any anatomical structure between the mitral valve base and the upper left pulmonary artery7 indicated complete surgical ligation of the left atrial appendage. Color Doppler flow identified incomplete ligation, which was shown as a flow crossing the separation between the left atrial body and the left atrial appendage.7 The contribution of left atrial appendage surgical obliteration to the risk of future embolism was analyzed using multivariate analysis.

	Fifty-two patients received complete ligation of the left atrial appendage, and six patients received incomplete ligation. Twenty-seven patients had an embolic event during the time between valve replacement and echocardiographic study (mean, 69.4 ± 67 months), of which 19 were ischemic strokes, five were peripheral embolisms, and three were transient ischemic attacks. In patients who did not receive left atrial appendage removal, systemic embolism was markedly increased (17%), compared with patients who underwent left atrial appendage ligation (3.4%; P = .01). Systemic embolism was also more common in patients without left atrial appendage ligation compared with those who received left atrial appendage ligation if left atrium or left atrial appendage thrombus (14.8% compared with 3.4%;

	P = .03) and left atrial appendage spontaneous echo contrast grade

	P < .001) were shown in the echocardiographic study (Table 1).

	The absence of left atrial appendage removal (odds ratio [OR], 6.7; 95% confidence interval [CI], 1.5—31.0; P = .02) and the finding of

	95% CI, 1.2—27.3; P = .03) were shown to be independent predictors of the occurrence of an embolic event after mitral valve replacement surgery on multivariate analysis (Table 2). The OR increased up to 11.9 (95% CI, 1.5–93.6; P = .02) when incomplete ligation shown by echocardiography was included in the model.

	It has been shown that the left atrial appendage is a cause of thromboembolic stroke. Stroke risk, therefore, would most likely be markedly diminished by removal of the left atrial appendage. Left atrial append-

	age ligation has been performed since early surgical treatment for rheumatic mitral stenosis to diminish embolization risk. The benefits of obliteration of the left atrial appendage during mitral valve replacement surgery are still being debated, however. It has been suggested that removing the left atrial appendage during cardiac surgery can help to prevent strokes.8 But there are few studies proving the advantages of this procedure.9,10 Although the procedure is endorsed in the American College of Cardiology guidelines, its use depends on the preferences and skill of the individual surgical team.11

	Incomplete surgical ligation of the left atrial appendage has been shown to have an important effect on the incidence of embolic events following surgery. In our study, 10.3% of patients had incomplete removal of the left atrial appendage and, of these, only one patient had an embolic event. It was shown in a recent study, however, that of 36%

	of patients with incomplete ligation of the left atrial appendage, 22% had an embolic event.7 The effect of the surgical technique is clear, and it can result in a greater incidence of postoperative embolic events.

	High-risk patients who received left atrial appendage removal during mitral valve replacement surgery were found to have a 6.7-fold reduced risk of embolism in our study. An additional decreased risk of embolism (11.9-fold) occurred when complete ligation was achieved and was shown on TEE. TEE may possibly be used to classify patients who are at greater risk for embolic events and therefore would receive greater advantage from ligation of the left atrial appendage. A greater risk of late embolism correlates with identification of the left atrial appendage or left atrium thrombus, higher left atrium size, and higher degrees of left atrial appendage spontaneous echo contrast on TEE.

	Because high-risk patients receive anticoagulants for many years, some physicians might believe that left atrial appendage ligation is not necessary. But eradicating the cause of embolic events would be a reasonable approach because warfarin (Coumadin) is often not used correctly or not used at all. In our study, we did not collect data on patients’ levels of anticoagulant medications. It may be possible that reduced in-ternational normalized ratios affected the development of an embo-

	lic event, despite the fact that all patients in our study who had a mechanical prosthesis received anticoagulant therapy. Regardless of the obvious clinical importance, we do not believe that the long-term anticoagulation profile is different for patients who undergo left atrial appendage ligation compared with those who do not receive left atrial appendage ligation. Therefore, in this setting, the amount of long-term anticoagulation therapy does not appear to be an important factor.

	Our study provides clear evidence for the benefit of obliteration of the left atrial appendage during surgical procedures as a method of preventing a postoperative embolic event and supports surgical left atrial appendage removal during mitral valve replacement. Until results of a blind, randomized study are available,12 the current study provides new data on the effects of left atrial appendage ligation.

Left atrial appendage obliteration for stroke reduction

	hest pain is one of the most common symptoms encountered in the emergency department.1 Although the characteristics of the pain, coronary artery disease (CAD) risk factors, and blood markers are all helpful in the evaluation of chest pain, a small number of patients with acute myocardial infarction (MI) are missed in the emergency department, and a large number of patients with low-risk chest pain are admitted for further evaluation, at considerable expense and inconvenience.2

	Relief of chest pain with nitroglycerin is often taken as a sign that the pain is cardiac in origin.3,4 Nitroglycerin has myriad effects, however, including relief of esophageal spasm,5 and thus the response to nitroglycerin may not have the diagnostic value that some attribute to it.

	We studied 459 patients who presented to the emergency department and were admitted with an initial diagnosis of “rule out MI,” “chest pain,” or the equivalent. They all had active chest pain in the ambulance or in the emergency department and had been given nitroglycerin. The pain level was recorded on a scale of 1 to 10 before and after administration of nitroglycerin. We then followed the patients through their hospitalization and contacted them by telephone at 4 months.

	We defined pain relief as a 50% or greater decrease in self-reported pain intensity following a single 0.4-mg dose of spray or sublingual nitroglycerin. The primary outcome was the presence of active CAD, which was defined as an increased serum troponin T level; 70% or greater stenosis on coronary angiogram; positive results on a stress test, with or without imaging; or typical symptoms in the setting of known severe disease in the absence of further testing.

	Of the 459 patients enrolled in the study, 181 (39%) had chest pain relief with nitroglycerin, and 278 (61%) did not. There were no significant differences in baseline characteristics between the two groups. Other medications given concurrently with the nitroglycerin were similar between the two groups. Additionally, it was shown that, in 141 patients (31%), the chest pain was caused by active CAD; however, 275 patients (60%) were shown not to have active CAD. For 43 patients (9%), we were not able to identify a definitive cause of the chest pain.

	Of the 141 patients whose chest pain was caused by active CAD, nitroglycerin provided relief in 49 (35%). For the 275 subjects whose chest pain was not due to active CAD, however, nitroglycerin provided chest pain relief in 113 patients (41%). These findings show that there was no diagnostic value to chest pain relief with nitroglycerin in the overall population. The Figure shows the sensitivity and specificity of chest pain relief by nitroglycerin (A; leftmost bars), along with the positive and negative likelihood ratios that chest pain would be relieved by nitroglycerin (B; leftmost bars). The likelihood ratios cross 1.0, which means that chest pain relief with nitroglycerin is not useful in differentiating active CAD from other diagnoses.

	We also examined the diagnostic value of chest pain relief in subgroups. As can be seen in the Figure, for patients who were shown not to have MI during the index admission (troponin negative), who had a history of CAD, and who did not have a history of CAD, the sensitivity and specificity of chest pain relief with nitroglycerin for active CAD were poor, and the likelihood ratios all crossed 1.0. This indicates that, even when the analysis was limited to these specific subgroups, chest pain relief by nitroglycerin was not use-ful in distinguishing active CAD from other sources of acute chest pain.

	We also examined other definitions of pain relief. By varying either  the percentage of pain change or an absolute reduction in pain level to define “pain relief,” we showed that no matter what definition was used, the likelihood ratios at each cutoff point were close to 1.0. This indicates that no matter what definition of pain relief was used, chest pain relief by nitroglycerin did not aid in the diagnosis of CAD.

	Although the diagnosis of chest pain was the primary outcome in this study, we also examined outcomes at 4 months to determine if chest pain relief by nitroglycerin held any prognostic value. As can be seen in the Table, there were no significant differences in the rates of death, MI, or revascularization between the patients who did and did not have pain relief (50% or greater reduction in self-reported pain) with nitroglycerin.

	The results of this study clearly show that chest pain relief with nitroglycerin does not aid in diagnosis and prognosis and should not be used in the triage of patients with chest pain who present to the emergency department. Our findings also held true when the analysis was limited to patients who did not have MI, had a known history of CAD, or had no known history of CAD.

	Pain relief also held no prognostic value in terms of 4-month outcomes. Even when the definition of pain relief was varied according to percentage of pain reduction or an absolute number of units of pain reduction, there was no diagnostic or prognostic value.

	Given that pain relief with nitroglycerin has been used as a criterion for CAD, how can our findings be explained? Nitroglycerin has many effects on the body, causing vasodilation of all vascular beds.6 It is a first-line treatment for esophageal spasm, which is high on the list of differential diagnoses for chest pain in patients in the emergency department.5 Additionally, there is probably a substantial placebo effect to nitroglycerin in the emergency department, as this is usually the first intervention given to patients with chest pain. Whether or not the pain is cardiac in origin, patients may report a decrease in pain intensity simply as an effect of the attention and any intervention.

	There are several limitations to our study. One is that we examined the effect of the first dose of nitroglycerin only. We chose this approach because the first nitroglycerin dose is usually given before other medications and interventions, such as oxygen, aspirin, and beta blocking agents. In many patients, the first nitroglycerin dose is insufficient to cause relief of pain, no matter what the cause, but by the second or third dose, the pain has been relieved. However, by this point, several other medications have typically been given, and distinguishing the effect of the nitroglycerin from the other interventions is difficult.

	Although our study was done in a prospective manner, it was observational. Additional cardiac testing, which formed the basis for deciding whether a patient had active CAD, was pursued at the option of the physicians caring for the patients after discharge and thus was not standardized. The pain was also based on the patient’s self-report using the simple 0- to 10-point scale. Although self-reported pain is not an ideal system,7 it does reflect clinical practice, and it was recorded in real time while the patients were having the pain; therefore, we believe that it accurately reflected the intensity of the pain.

	The study was conducted solely with patients who presented to the emergency department with chest pain. Although our results strongly support the conclusion that chest pain relief with nitroglycerin holds little diagnostic or prognostic value in the emergency department setting, they do not address chest pain relief from nitroglycerin in other settings. We suspect that there is substantial overlap between patients with chest pain presenting to the emergency department and those presenting to physicians’ offices and that chest pain relief with nitroglycerin is most likely not useful in this setting, either.

	In the emergency department setting, chest pain relief from nitroglycerin does not predict active CAD and does not predict outcomes at 4 months. Response of chest pain to nitroglycerin should not be used to guide triage in the emergency department and does not yield any information about the underlying disease process.

Can chest pain  relief by nitroglycerin  predict active CAD?

	The prevalence of diabetes has been increasing for the past several decades. The total number of people with diabetes in the world is projected to increase from 171 million in the year 2000 to 366 million in 2030,1 with more than 90% of them having type 2 diabetes. Both type 2 diabetes and its precursor, impaired glucose tolerance, increase with age. In the United States, approximately 20% of people 60 years of age or older have type 2 diabetes, and another 15% have impaired fasting glucose,2 which is an underestimate of impaired glucose tolerance. The risk is even higher in certain ethnic populations, such as Hispanic, African American, and Asian American.3 Diabetes also has a substantial effect on morbidity and mortality.

	Although the microvascular complications of diabetes are a reflection of the severity of hyperglycemia the patient has experienced over the years, the cardiovascular complications are not closely related to the duration or extent of hyperglycemia. Several large epidemiologic studies, such as the Diabetes Epidemiology Collaborative Analysis of Diagnostic Criteria in Europe study, have shown a 1.5- to twofold increased risk of cardiovascular disease and mortality in people with impaired glucose tolerance and a relatively weaker association with impaired fasting glucose.4 The increased risk of cardiovascular disease events and mortality may be seen in the presence of impaired glucose tolerance, even when it does not progress to diabetes on long-term follow-up5 (as documented in the FINMONICA stroke register, a Finnish subset of the World Health Organization [WHO] MONItoring of trends and determinants of CArdiovascular disease study). These observations support the role of underlying atherogenic risk factors in the presence of impaired glucose tolerance. Some of the clinically recognizable risk factors frequently accompanying “prediabetes” (impaired glucose tolerance or impaired fasting glucose) include hypertension, central obesity, insulin resistance, microalbuminuria, and dyslipidemia, collectively now well recognized as the metabolic syndrome by the National Cholesterol Education Program Adult Treatment Panel III, WHO, the American Association of Clinical Endocrinologists, and other organizations. Recently published cross-sectional analyses of more than 10,000 subjects in the Third National Health and Nutrition Examination Survey (NHANES III)6 and a meta-analysis of 11 prospective European studies, including more than 11,500 subjects,7 have confirmed a significant predictability of the multiple components of the metabolic syndrome for cardiovascular disease events or mortality.

	Although the presence of impaired fasting glucose is an infrequent component of the metabolic syndrome,8 impaired glucose tolerance is more often seen in patients with cardiovascular disease or cardiovascular disease—associated risk factors,9 in keeping with the epidemiologic observations.4-7 Recently, several prospective studies have addressed the issue of the prevalence of diabetes in patients with established cardiovascular disease and in those at high risk for cardiovascular disease events (Table). Norhammar and colleagues evaluated a group of 181 patients (mean age, 63.5 years) following acute myocardial infarction (MI) and discovered that, 3 months after MI, 25% of those patients had previously undiagnosed diabetes and an additional 40% had impaired glucose tolerance, as shown by oral glucose tolerance tests.10 Thus, a total of 65% of those subjects had evidence of abnormal glucose homeostasis, of which more than 50% would have been missed by fasting blood glucose alone.

	Two recent reports are from the largest studies of glucose tolerance tests in patients with cardiovascular disease or at high risk for cardiovascular disease. The first was a prospective, observational study in 25 European countries, the Euro Heart survey, involving 4,961 patients with acute or stable coronary artery disease.11 Of these, 31% had previously diagnosed diabetes mellitus. However, on oral glucose tolerance testing of approximately 1,900 patients without known diabetes, about 4% had impaired fasting glucose, 32% had impaired glucose tolerance, 18% had diabetes mellitus, and 45% had normal glucose tolerance. The second study was an even larger multinational survey performed during the screening of eligible patients for the ongoing Nateglinide and Valsartan in Impaired Glucose Tolerance Outcomes Research (NAVIGATOR) trial.12 In this survey, 43,509 nondiabetic subjects (mean age, 63 years; mean body mass index, 29.7 kg/m2; 49% men, 51% women) were screened for a diagnosis of impaired glucose tolerance. Of these, 9,125 had evidence of prior cardiovascular disease. In addition, 31,047 had no prior cardiovascular disease but had one or more risk factors for cardiovascular disease, including positive family history of premature cardiovascular disease, smoking, hypertension, low-density lipoprotein cho-lesterol level above 160 mg/dL, high-density lipoprotein cholesterol level below 38 mg/dL, triglyceride level above 200 mg/dL, left ventricular hypertrophy, or microalbuminuria. Among those with prior cardiovascular disease, the proportion of patients with normal glucose tolerance, impaired fasting glucose, impaired glucose tolerance, and diabetes mellitus were 34%, 12%, 29%, and 24%, respectively. Similarly, among those without prior cardiovascular disease but with one or more risk factors, the distribution

	of these categories was 38%, 12%, 29%, and 21%, respectively, virtually identical to the group with previous evidence of cardiovascular disease. Thus, this rather large survey confirmed that approximately one of five people had previously undiagnosed diabetes, similar to the expected number in this age group, but an additional one of four had other abnormalities of glucose tolerance (prediabetes). Of note, impaired glucose tolerance was more than twice as common as impaired fasting glucose in those with prediabetes (impaired glucose tolerance and impaired fasting glucose).

	This major database of the ongoing NAVIGATOR trial underscores the preponderance and significance of prediabetes and undiagnosed diabetes in patients with cardiovascular disease. In addition, there was an impressive similar frequency of these disorders, even for those who may have been considered to be in a “pre-

	cardiovascular” disease state in view of the presence of several cardiovascular disease risk factors, which accounted for the 51% prevalence of metabolic syndrome in this population. Furthermore, it underscores the need for earlier diagnosis of impaired glucose tolerance in this population which, when detected, will provide an opportunity to possibly prevent progression to diabetes and to intervene with earlier detection and management of cardiovascular disease risk factors. Finally, these observations further support the likelihood of common antecedents (eg, endothelial dysfunction, subclinical inflammation, and augmented sympathetic and/or renin-angiotensin system activity) in the pathogenesis of type 2 diabetes as well as cardiovascular disease.

	In addition to practical lifestyle measures that have been shown to reduce the progression to diabetes, a number of pharmacologic agents are being evaluated for their potential efficacy in diabetes prevention. These include metformin (Glucophage), thiazolidinediones, alpha-glucosidase inhibitors, and anti-obesity drugs, such as orlistat (Xenical). Of great interest is the possibility that, based on the subanalyses from a number of recent trials, drugs that antagonize the renin-angiotensin system, such as angiotensin-converting enzyme inhibitors,13 angiotensin receptor blockers,14-16 and even certain calcium channel blocking agents,17 may have significant benefits in diabetes prevention (ie, a 20%—30% reduction). This is the subject of ongoing trials, including the Diabetes Reduction Approaches with Ramipril and Rosiglitazone Medications and NAVIGATOR studies, which examine this possibility in an a priori design with such agents.

Diabetes and Cprediabetes 201D  in cardiovascular disease

	Over the past several decades, clinical research in cardiology has been driven, in part, by large, multicenter trials. These in turn have generated large clinical data sets, which have created both opportunities and challenges for clinical researchers. The opportunities lie in the ability to mine these data sets to establish correlations among clinical variables; the challenges lie in generating plausible hypotheses that provide insight into the pathophysiology of the underlying disease and strategies for novel treatments.

	The current study by Anand and colleagues (page 12) is an example of both the strengths and limitations of this strategy. The investigators analyzed a group of 912 subjects with congestive heart failure (CHF) who were originally enrolled at multiple sites in a trial investigating the effect of a recombinant human tumor necrosis factor receptor that binds to and inactivates the soluble cytokine.1 The study was designed and powered to detect differences in the hard end points of all-cause mortality and repeat hospitalization and was terminated prematurely when it became clear that no differences were evident.

	The 912 subjects subsequently provided a fertile data pool, and the present study derives from a posthoc analysis of clinical variables that correlated with adverse outcomes in this group, regardless of treatment strategy. This analysis strongly suggests that anemia is common in patients with moderately severe CHF (hemoglobin levels of less than 12.5 g/dL were seen in 20% of the cohort) and that the magnitude of the anemia is associated with greater disease severity, greater left ventricular mass (defined in a very small group of 66 subjects who underwent cardiac magnetic resonance imaging), and worse clinical outcomes. In a multivariate analysis, anemia appeared to be an independent predictor of outcome when the data were adjusted for intercurrent pharmacotherapy, New York Heart Association functional class, diastolic blood pressure, and creatinine (albeit at P = .021). While this is a powerful observation, it is certainly appropriate to caution that several other potentially important variables were not comprehensively evaluated, including age, sex, etiology of heart failure, and adequacy of pharmacotherapy, among others.

	This observation that anemia is seen in heart failure patients and heralds poor outcomes is not novel to this analysis; it has been seen previously in other studies2-5 and is intuitively appreciated by heart failure physicians. What the finding tells us about the natural history of CHF, however, and whether correction of the anemia has the potential to affect this natural history is not answered by these studies, and is not at all clear.

	Because the anemia is generally characterized as being reflective of a chronic disease, the finding may be a very indirect marker of disease severity or impaired renal perfusion and/or intercurrent renal disease. Alternately, the anemia may limit substrate delivery to skeletal muscle and amplify symptoms, tax the cardiovascular system, and accelerate the progression of the disease. Some data in support of this latter concept come from the work of Silverberg et al6 and Mancini et al,7 who showed that erythropoietin improved exercise capacity in small groups of patients with moderate to severe heart failure. Until we have more information about the reasons for the anemia and its contribution to disease progression, however, attempts to correct it with the expectation that we can alter disease progression seem premature.

How common is anemia  in heart failure?

	A 76-year-old man was admitted to the hospital and diagnosed with stroke caused by peripheral embolism. He underwent cardiac surgery 12 years earlier for replacement of mitral and aortic valves with biological prostheses. During surgery, left atrial appendage ligation was not performed and, although the prostheses implanted were biological, the patient received anticoagulant treatment with warfarin.

	A computed tomography scan showed that the patient had a stroke, and transesophageal echocardiography (TEE) was performed to search for an embolic source. The echocardiographic study showed normal function of both prostheses and an enlarged left atrial appendage with spontaneous echo contrast and low appendage flow velocities. In addition, a thrombus located inside the left atrial appendage was found in the TEE study (Figure).

	This patient developed a thrombus in the left atrial appendage despite treatment with anticoagulant therapy. This case illustrates a high-risk patient who would probably have benefited from surgical left atrial appendage li-gation during cardiac surgery. The absence of left atrial appendage ligation together with a low-flow velocity and spontaneous echo contrast are variables that, according to the findings of our study, are associated with increased stroke risk. n

Case report:  A stroke patient with  left atrial appendage

	A 65-year-old man with ischemic cardiomyopathy and symptoms of worsening heart failure was examined in the cardiology outpatient clinic. The patient had been diagnosed with heart failure 6 years earlier and was receiving maximally tolerated doses of a diuretic, an angiotensin-converting enzyme inhibitor, and a beta blocking agent. His ejection fraction was 33%, and his hemoglobin concentration was 11.6 g/dL. Iron/transferrin saturation and B12, folate, and ferritin concentrations were all normal.

	On examination, the patient was shown to be in New York Heart Association functional class 3, with a peak oxygen consumption (VO2) of 12.2 mL/kg/min (mean of two assessments performed 2 weeks apart), a maximal exercise duration of 482 seconds (assessed using a modified Naughton protocol), and a serum brain natriuretic peptide (BNP) concentration of 413 pg/mL. The patient entered a double-blind, randomized, placebo-controlled clinical trial examining the effect of hemoglobin optimization on functional status in chronic heart failure patients. He was randomly assigned to receive active treatment with darbepoetin alfa (Aranesp). Darbepoetin alfa is an erythropoiesis-stimulating protein that can be administered once every 2 weeks or once a month. It is currently only approved for the treatment of anemia associated with chronic renal failure and chemotherapy-induced anemia in patients with nonmyeloid malignancy.

	Darbepoetin alfa was initiated at a dose of 0.75 µg/kg, administered by subcutaneous injection every 2 weeks. Hemoglobin concentration was monitored weekly. The dose of darbepoetin alfa was increased by 25% every 4 weeks to achieve an increase in hemoglobin concentration of between 0.5 and 1.5 g/dL every 3 weeks and a target hemoglobin concentration of 14 g/dL. The patient reached the target hemoglobin concentration after 13 weeks of treatment. The hemoglobin concentration was 13.8 g/dL 27 weeks after initiating treatment, and the patient’s functional status was both subjectively and objectively improved. Peak VO2 had increased by 2.2 mL/kg/min, to 14.4 mL/kg/min. Treadmill exercise duration had improved from 175 seconds to 657 seconds, and the patient’s BNP concentration had fallen by 77 pg/mL, to 336 pg/mL. At week 27, the patient was receiving 0.85 µg/kg of darbepoetin alfa every 2 weeks.

Case report: Anemia  in a patient with chronic heart failure

	Anemia can be a powerful independent risk factor for mortality in chronic heart failure patients with both systolic and diastolic dysfunction.1-4 We studied a group of patients enrolled in the Randomized Etanercept North American Strategy to Study Antagonism of Cytokines (RENAISSANCE) trial, the details of which have been published elsewhere,5 to analyze the relationship between anemia, severity of chronic heart failure, and clinical outcome. We also evaluated the correlation between hemoglobin concentration and left ventricular mass.

	The RENAISSANCE trial was a double-blind multicenter study that included 912 patients with moderate to severe chronic heart failure who were randomly assigned to receive etanercept (Enbrel, a tumor necrosis factor-a antagonist) or placebo. Blood samples were taken at baseline and at 2, 4, 12, and 24 weeks. A subgroup of patients underwent magnetic resonance imaging (MRI) at the beginning of the study and at 24 weeks. For MRI studies and hemoglobin concentrations that were missing after baseline visits, a last-observation-carried-forward approach was employed. To determine whether changes from the baseline were different from 0, the Wilcoxon signed rank test was used. The Wilcoxon rank sum test was used to compare changes from baseline among groups. The log-rank test was used to perform analyses of the end points of time to death and time to death or chronic heart failure hospitalization. Continuous predictor variables were analyzed using the Cox proportional hazards regression model. P values below .05 were considered to show statistical significance.

	Of all patients enrolled in the RENAISSANCE study, 12% were anem-

	ic (when anemia was defined as a hemoglobin concentration ≤ 12.0 g/dL). When anemia was defined as a hemoglobin concentration of 12.5 g/dL

	or lower, 20% of subjects were anemic. With increased worsening of New York Heart Association (NYHA) functional class, the rate of anemia increased.

	Sixty-six subjects had both he-moglobin concentrations and MRI scans taken. Over the 24 weeks of the study, when hemoglobin concentrations decreased or stayed the same, left ventricular mass index increased by a mean of 1.6 g/m2

	(± 7.9 g/m2 SD; Figure 1). Left ventricular mass index decreased by 7.5 g/m2 (± 12.1 g/m2 SD), however, for subjects whose hemoglobin concentrations increased (P = .008). Based on these results, it appears that a

	4.1-g/m2 reduction in left ventricular mass index occurred for every

	1.0-g/dL increase in hemoglobin concentration.

	Nonanemic patients (hemoglobin levels > 12.0 g/dL) had a low-er mortality rate than anemic patients (16% compared with 28%, respectively; P = .0172; log-rank test). Mortality and chronic heart failure hospitalization rates were also lower for nonanemic patients (33% compared with 56%, respectively; P < .001; log-rank test; Figure 2). A 15.8% decrease in the risk of death and a 14.2% decrease in the risk of death and chronic heart failure hospitalization occurred for every 1-g/dL increase in hemoglobin concentration. Patients with anemia had higher event rates for the two end points of death and death and chronic heart failure hospitalization irrespective of NYHA functional class (Figure 3), and mortality increased in a hemoglobin-dependent manner (Figure 4). Baseline hemoglobin concentration was an independent predictor of mortality or chronic heart failure hospitalization (hazard ratio = 0.915; P = .021), even after controlling for NYHA functional class, serum creatinine concentration, baseline diastolic blood pressure, and use of antiarrhythmic drugs and beta blocking agents. Lower mortality correlated with a higher hemoglobin concentration (Figure 4). A U-shaped association was not found. In a Kaplan-Meier analysis, 97 of the 466 subjects whose hemoglobin concentrations decreased or stayed the same died (18.1%) as compared with only 62 of 384 subjects (11.4%, P = .008; log-rank test) whose hemoglobin concentrations increased during the same period.

	The causes of anemia in patients with chronic heart failure, commonly called anemia of chronic

	disease, are not clear. Low levels of erythropoietin caused by angiotensin-converting enzyme inhibitors and other drugs, the effects of cytokines, and decreased renal function are considered possible causes; low levels of hematinics, however, do not seem to be involved.6

	A few studies have shown that treating anemia in chronic heart failure patients can be advantageous. Marked improvements in visual analogue score, ejection fraction, hospital events, and NYHA functional class were shown in a study in which anemia was treated with erythropoietin in chronic heart failure patients.7 Results of another study showed marked improvements in exercise duration, performance in the 6-minute walk test, peak oxygen consumption, and quality of life.8 Erythropoiesis-stimulating proteins and blood transfusions can be used to correct anemia in patients with chronic heart failure. However, none of the erythropoiesis-stimulating proteins (epoetin alfa [Epogen, Procrit]; epoetin beta; and darbepoetin alfa [Aranesp]) are indicated for anemia in chronic heart failure patients. Darbepoetin alfa can be given subcutaneously every 2 weeks or every month because of its half-life.

	A number of placebo-controlled randomized studies are being conducted using darbepoetin alfa. Others are being planned to determine whether correction of anemia (defined as a hemoglobin concentration ≤ 12.5 g/dL) in chronic heart failure patients results in decreased morbidity and mortality and improved functional status.

Anemia and clinical outcome in heart failure

	ORLANDO—Cardiac resynchronization therapy (CRT) reduces death or hospitalization due to a major cardiovascular event by 37% in patients with advanced heart failure, according to results of the Cardiac Resynchronization in Heart Failure (CARE-HF) trial. “This is the first study to show a survival benefit with pacemakers,” said John G.F. Cleland, University of Hull, United Kingdom. “Other shorter trials have shown only that CRT improved 6-minute walk times and quality of life in these patients.”

	The addition of CRT to medical therapy also significantly lowered New York Heart Association (NYHA) functional heart failure class and significantly increased ejection fraction. “By 18 months a large proportion of CRT patients became asymptomatic with ejection fractions over 40%,” said Dr. Cleland. “We can now begin to talk about remission of heart failure in 15% to 20% of patients.”

	Dr. Cleland reported the CARE-HF results during the American College of Cardiology Annual Scientific Sessions. A total of 813 patients with class III or IV heart failure from 82 international centers received either optimal medical therapy (n = 404) or CRT implantation plus medical therapy (n = 409). Eligible patients had heart failure of at least 6 weeks’ duration requiring loop diuretics despite optimal pharmacologic therapy, left ventricular systolic dysfunction and dilation with an ejection fraction of 35% or less, and QRS intervals of 120 ms

	or greater or QRS between 120 and 149 ms with dyssynchrony confirmed by echocardiogram. Patients with atrial fibrillation or those who required pacing were excluded.

	Average patient age was 66.5 years, 73.5% were male, the majority had class III heart failure, and 38% had ischemic heart disease. A majority of patients in both groups were treated with angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, beta blocking agents, and spironolactone. Fewer than half the patients in each group were treated with digitalis (Lanoxin) and less than half took furosemide (Lasix) in doses of greater than 80 mg daily. CRT patients received an InSync-III device. The mean follow-up was 29.4 months.

	The primary end point was a composite of all-cause mortality or an unplanned hospitalization for a major cardiovascular event. After 1,300 days, significantly fewer CRT patients reached the primary end point compared to those on pharmacologic therapy alone (P < .0001, hazard ratio [HR] 0.63). All-cause mortality alone was also lower among CRT patients (P = .0019, HR 0.64), amounting to a 36% reduction compared with medical therapy patients.

	After 90 days, NYHA class was 2.7 in the medical therapy alone group compared with 2.1 in the

	CRT group (P < .0001). Quality-of-life scores were also significantly better among CRT patients (P < .0001). Dr. Cleland said that after 18 months CRT patients exhibited highly significant differences compared with the control group in terms of shorter interventricular mechanical delay, higher ejection fraction by about 7%, less mitral regurgitation, lower ventricular volumes, higher systolic blood pressure, and lower N-terminal-pro-brain natriuretic peptide levels.

	“CRT should be considered routine therapy for patients with moderate to severe heart failure with evidence of cardiac dyssynchrony if their symptoms persist despite optimal drug therapy in order to improve cardiac function and efficiency, improve symptoms and quality of life, reduce morbidity, and prolong survival,” Dr. Cleland said. He added that CARE-HF is especially important owing to the length of follow-up compared with previous CRT trials. n

CRT proven to reduce heart failure mortality

	ORLANDO—Aggressive lipid lowering using 80 mg/day of atorva-statin (Lipitor) prevents more major adverse cardiovascular events than 10 mg/day of atorvastatin in patients with stable coronary heart disease (CHD), said John LaRosa, MD, at the American College of Cardiology Annual Scientific Sessions 2005.

	In reporting the results of the Treating to New Targets (TNT) study, Dr. LaRosa said that the reduction in cardiovascular risk with high- versus low-dose atorva-statin was “very robust” but that new guideline recommendations for lowering low-density lipoprotein (LDL) cholesterol for patients with stable CHD should await completion of other studies in progress. He did offer that the “link between LDL cholesterol and coronary heart disease is no longer disputed.”

	In TNT, 10,001 patients with stable CHD and LDL cholesterol values of less than 130 mg/dL were randomly assigned to 10 or 80 mg/day of atorvastatin. The average exposure to treatment was 4.9 years, at which time patients assigned to 10 mg/day of atorvastatin achieved a mean LDL cholesterol level of 101 mg/dL and those assigned to 80 mg/day of atorvastatin achieved a mean LDL cholesterol of 77 mg/dL.

	The primary end point—a composite of death from CHD, nonfatal myocardial infarction, resuscitation from cardiac arrest, or fatal or nonfatal stroke&mdash;was reduced by 22% (P = .0002) with high- versus low-dose atorvastatin. The outcome of stroke was reduced by 25% in the 80-mg/day group (P = .02).

	There was no significant difference in mortality between the two groups, but the study was not sufficiently powered to compare the effect of the two doses on overall mortality, explained Dr. LaRosa, president, State University of New York Downstate Medical Center, Brooklyn. “We cannot with 10,000 people even answer the issue of mortality, and this is going to be a problem from now on in clinical trials in this area,” he said. “We have to make the assumption…that lowering LDL does lower the total mortality rate.”

	The incidence of persistent elevations in liver enzyme levels was 0.2% in the group randomly assigned to 10 mg/day and 1.2% in the group assigned to 80 mg/day of atorvastatin (P < .001), “which is a fraction of the rate in other statin trials,” he said.

	The incidence of persistent elevations in creatine kinase levels or complaints of myalgia did not differ between the two groups. Only five cases of rhabdomyolysis were reported, none of which the investigators attributed to the study drug.

	Any change in the target levels of LDL cholesterol for patients with stable CHD should await the results of additional studies in which aggressive and moderate lipid lowering are compared, which are expected to be completed soon, said Dr. LaRosa. “We would be on firmer ground if we waited,” he said. “We already have an option to go to 70 [mg/dL].” In 2004, the Adult Treatment Panel (ATP) of the National Cholesterol Education Program issued an update to ATP III guidelines in which an option to achieve a more aggressive LDL cholesterol—70 mg/dL&mdash;was offered at the discretion of the treating physician.

High-dose better than low-dose atorvastatin in preventing events in stable CHD

	ORLANDO—Results of the Optimal Pharmacological Therapy in Implantable Cardioverter Defibrillator Patients (OPTIC) trial show that the use of amiodarone (Cordarone, Pacerone) with a beta blocking agent dramatically reduces shocks from an implantable cardioverter defibrillator (ICD) by 73% compared to the use of sotalol (Betapace) or a beta blocker alone, reported Stuart J. Connolly, MD, McMaster University, Hamilton, Ontario, Canada. “The results represent a huge reduction in shocks and a big improvement in quality of life,” he said.

	“Why should we devote so much attention to reducing shocks? Because ICD patients routinely report anxiety, depression, fear of the device firing, and fear of death when the device fires,” Dr. Connolly said during the American College of Cardiology Annual Scientific Sessions. For example, a recent Mayo Clinic study reported that 24% of ICD patients have daily anxiety related to the device. Two years postimplantation, 14% of patients continue to report a daily preoccupation with their ICD.

	In the 3-year OPTIC trial, patients from 39 international centers receiving a dual-chamber ICD were randomly assigned to receive a beta blocker alone (n = 136), amiodarone plus a beta blocker (n = 140), or sotalol (n = 134). Patients had documented sustained ventricular tachycardia (VT) or ventricular fibrillation (VF)/cardiac arrest with ejection fractions of 40% or less, or inducible VT or VF with either ejection fractions of 40% or less or recent unexplained syncope. The majority of patients had spontaneous VT or VF and a history of myocardial infarction. Those with long QT intervals, previous chronic amiodarone or sotalol treatment, or a need for a class I or III antiarrhythmic agent were excluded.

	Beta blockers used included metoprolol (Lopressor, Toprol) at a target dose of 100 mg daily, bisoprolol (Zabeta, Ziac) at 10 mg, or carvedilol (Coreg) at 50 mg daily. Patients receiving a beta blocker plus amiodarone took 800 mg amiodarone daily for the first 2 weeks, 400 mg daily for an additional 4 weeks, and 200 mg daily thereafter. Patients in the sotalol group received 240 mg per day in two or three divided doses or 160 mg daily if creatinine clearance was 30 to 60 mL/minute. Baseline characteristics were similar for each treatment arm.

	ICDs were optimally programmed for antitachycardia pacing therapy for VT episodes less than 222 beats per minute with four bursts of 10 beats, an initial burst of 85% of VT cycle length, and 12-ms decrements. Supraventricular tachycardia discrimination was turned on to detect onset, stability, and morphology with the timer set to 2 minutes.

	The risk of any ICD shock dur-ing the first year after implantation was dramatically lower among patients in the amiodarone and beta blocker group compared with those taking a beta blocker alone or sotalol. “Sotalol was only somewhat effective in reducing shocks,” Dr. Connolly said.

	Overall, the annual shock rate was 38.5% for beta blocker alone, 10.3% for amiodarone plus beta blocker therapy, and 24.3% for sotalol. The annual rate of appropriate ICD shocks was 22.0%, 6.7%, and 15.1%, respectively; and the annual rate of inappropriate ICD shocks was 15.4%, 3.3%, and 9.4%, respectively. The mean number of shocks per patient during follow-up was 4.32 among beta blocker patients, 0.51 in the amiodarone and beta blocker group, and 0.93 among sotalol patients. Amiodarone and beta blocker treatment reduced appropriate, inappropriate, and frequent shocks. “In the future, additional studies to measure whether reduced anxiety levels are associated with less frequent shocks will be important,” Dr. Connolly said.

	Amiodarone combined with a beta blocker was also better tolerated, with 18% of patients discontinuing therapy compared with 24% of sotalol patients who discontinued therapy. Adverse events were lower overall, however, in the beta blocker alone group. Atrial fibrillation occurred in 4.4% of beta blocker patients, 0.7% of amiodarone and beta blocker patients, and in 4.5% of sotalol patients. Pulmonary adverse events occurred in 5.0% of amiodarone and 3.0% of sotalol patients and did not occur in those taking beta blockers alone. Similarly, hypo- or hyperthyroidism occurred in 5.7% of amiodarone and 0.8% of sotalol patients but in none of the beta blocker alone patients. Symptomatic bradycardia occurred in 0.7% of beta blocker patients compared with 6.4% of amiodarone and 1.5% of sotalol patients.

April 2005

Chest pain relief from nitroglycerin: Very little comfort

Case report: Chest pain relief by nitroglycerin

Left atrial appendage obliteration for stroke reduction

Can chest pain relief by nitroglycerin predict active CAD?

Diabetes and Cprediabetes 201D in cardiovascular disease

How common is anemia in heart failure?

Case report: A stroke patient with left atrial appendage

Case report: Anemia in a patient with chronic heart failure

Anemia and clinical outcome in heart failure

CRT proven to reduce heart failure mortality

High-dose better than low-dose atorvastatin in preventing events in stable CHD

Amiodarone drastically reduces ICD shocks

Latest News