Prognostic Significance of Plasma Norepinephrine in Patients With Asymptomatic Left Ventricular Dysfunction
Background Elevated plasma neurohormonal levels are associated with increased mortality rates in patients with symptomatic heart failure. A previous Studies of Left Ventricular Dysfunction (SOLVD) trial suggested that neurohumoral activation precedes the development of symptoms as demonstrated by increased neurohormonal levels in patients with asymptomatic left ventricular dysfunction. However, the significance of this early neurohumoral activation is unclear. The goals of this study were to determine the prognostic significance of the plasma concentrations of plasma norepinephrine (PNE) and atrial natriuretic peptide (ANP) and the renin activity (PRA) in patients with asymptomatic left ventricular dysfunction.
Methods and Results PNE and PRA were measured before randomization in 514 patients with left ventricular ejection fractions ≤35% who did not require treatment for congestive heart failure and were enrolled in the SOLVD Prevention Trial. Plasma ANP levels were measured in a subset of 241 patients owing to study design. Using the Cox proportional hazards model that included left ventricular ejection fraction, New York Heart Association functional class, age, sex, treatment assignment to placebo or enalapril, and cause of heart failure, we examined whether these neurohormones predicted all-cause mortality, cardiovascular mortality, hospitalization for heart failure, development of heart failure, or development of ischemic events (myocardial infarction or unstable angina). PNE was the strongest predictor of clinical events in this patient population. PNE levels above the median of 393 pg/mL were associated with a relative risk of 2.59 (P=.002) for all-cause mortality, 2.55 (P=.003) for cardiovascular mortality, 2.55 (P=.005) for hospitalization for heart failure, 1.88 (P=.002) for development of heart failure, 1.92 (P=.001) for ischemic events, and 2.59 (P=.005) for myocardial infarction. PNE remained the most powerful predictor for all-cause mortality and ischemic events when the analysis included only the patients with histories of ischemic left ventricular dysfunction. The increases in other neurohormonal levels were not useful in predicting the subsequent development of clinical events.
Conclusions Increased PNE levels in patients with asymptomatic left ventricular dysfunction appear to predict all-cause and cardiovascular mortalities and development of clinical events related to the onset of heart failure or acute ischemic syndromes. Thus, measurement of PNE may be a possible early marker for assessment of disease progression in patients with left ventricular dysfunction, and modulating the release or effect of PNE may lead to improved prognosis and/or a reduction in morbidity.
The concentrations of several neurohormones are increased in plasma in patients with CHF, including PNE,1 2 ANP,3 and PRA.4 As clinical symptoms of heart failure advance, neurohormonal activation becomes more marked.2 5 6 These elevated levels of neurohormones are associated with increased mortality rates in patients with chronic CHF. For example, marked increases in PNE,5 PRA,7 or ANP8 levels are associated with increased mortality rates in patients with advanced symptoms of CHF. However, it is unclear whether the elevation of neurohormones is a late manifestation secondary to the severity of heart failure or an early marker that precedes the clinical syndrome and predicts subsequent outcomes.
The significance of neurohormonal activation may be better studied in patients with asymptomatic LV dysfunction with subsequent follow-up until clinical heart failure develops. A previous SOLVD trial indicated that patients with asymptomatic LV dysfunction had early increases in PNE, PRA, and ANP compared with normal control subjects.9 However, it is unknown whether the elevated neurohormonal levels in these patients can predict increased mortality, hospitalization for CHF-related symptoms, or development of ischemic events (myocardial infarction or unstable angina). Therefore, in the SOLVD Prevention Trial, we determined whether the neurohormonal levels measured before randomization were predictive of mortality and the development of clinical events related to the development of heart failure and acute ischemic events in patients with asymptomatic LV dysfunction.
SOLVD, a large clinical trial involving patients with LV dysfunction and heart failure, was conducted in 23 medical centers in the United States, Canada, and Belgium. It included two main trials and a registry. In the two trials, patients were randomized to receive enalapril or placebo for an average of 40 months. Details of the protocol, including patient selection, randomization, and follow-up, were published previously.10 All the participants enrolled in the Prevention Trial (asymptomatic or minimally symptomatic LV dysfunction with ejection fraction ≤35%) who did not require treatment for CHF and in whom a prerandomization neurohormonal measurement was available were included in this analysis. SOLVD included several substudies in which neurohormonal levels were measured with a common protocol at baseline. These included the neurohumoral (n=204), echocardiogram–sudden death (n=192), and radionuclide (n=118) substudies. Descriptions of these substudy designs are provided elsewhere.10
The prognostic usefulness of the neurohormones was examined by use of the Cox proportional hazards model. We looked at the relation between the prerandomization PNE, ANP, and PRA levels and all-cause mortality, development of CHF or hospitalization for CHF symptoms, and occurrence of ischemic events (development of myocardial infarction or unstable angina). The definitions of these clinical events used in SOLVD were given previously.10 Development of these clinical events was assessed to the day of death or hospitalization for CHF symptoms, to the day of hospitalization for ischemic events, to the day of the clinic visit at which the participant was diagnosed as having developed CHF symptoms, or to the end of the study on April 1, 1991. The mean duration of follow-up was 916 days.
The study protocols were approved by the local hospital review boards and the NHLBI. All patients provided written informed consent to participate in the study.
An intravenous cannula was inserted into an arm vein and flushed with heparinized saline. Participants were allowed to rest in the supine position in a quiet room for 30 minutes. Blood (15 mL) was collected for analysis of PNE, PRA, and ANP. The details of sample collection and processing and the quality control method used are given elsewhere.11 Samples from all the centers were shipped on dry ice for analysis by the SOLVD Neurohormone Core Laboratory at the University of Texas Medical School at Houston. PNE was measured by a radioenzymatic method. This assay for PNE has a sensitivity of 1 pg/mL with an interassay coefficient of variation of 6.1%.12 PRA was measured with the modified method of Sealey and Laragh.13 This assay for PRA has a sensitivity of 2 pg/mL of angiotensin I (0.1 ng·mL−1·h−1 PRA) with an interassay coefficient of variation of 12.6%.13 Plasma ANP was measured by a simplified radioimmunoassay. This assay for ANP has a sensitivity of 2 pg/mL with an interassay coefficient of variation of 13.7%.14 All samples were analyzed by “blinded” investigators who had no knowledge of the patient characteristics.
The Wilcoxon rank sum and χ2 tests were used to compare the continuous and categorical covariates between the neurohormones study group (n=514) and the remaining Prevention Trial patients (n=4228−514=3714). The Cox proportional hazards model used for the analysis of prerandomization neurohormonal levels adjusted for the following variables: age, sex, ejection fraction, NYHA functional class, cause of heart failure, and assignment to placebo or enalapril. Data are reported as median and interquartile range, and all comparisons are two sided. The reported risk ratios are multivariate-adjusted risk ratios. The data for death or hospitalization for heart failure symptoms also are presented as a combined end point to avoid the problem of competing risks.15
Baseline Clinical Data
Of the 514 patients in this study, prerandomization PNE and PRA levels were available in 509 (99%) and 512 (99%) participants, respectively, enrolled in the Prevention Trial. Because of differences in substudy design, prerandomization ANP levels were measured in only 241 participants. Table 1⇓ gives the clinical characteristics of these participants. These characteristics were not significantly different from those of the participants enrolled in the main Prevention Trial population described elsewhere.16 As in the Prevention Trial, most of the participants were men, and ischemic heart disease was the predominant cause of CHF.
PNE Levels and Development of Clinical Events
The median PNE level before randomization was 393 pg/mL, with an interquartile range of 291 to 528 pg/mL (439±224 pg/mL, mean±SD) in the 509 participants. In participants with PNE levels greater than the median value (n=254), there were 42 deaths (16.5%). In participants with PNE levels below the median (n=255), there were only 15 deaths (5.9%). Fig 1A⇓ shows adjusted survival curves for these two groups (risk ratio, 2.59; 95% CI, 1.42 to 4.70; P=.002; Fig 2⇓ and Table 2⇓). The median lengths of survival were 930 and 976 days in those participants with PNE levels above and below the median value, respectively. This difference in mortality was due almost entirely to an increase in deaths resulting from cardiovascular causes (39 versus 14 in the groups with PNE above and below the median, respectively; risk ratio, 2.55; 95% CI, 1.37 to 4.74; P=.003; Fig 1B⇓). Table 2⇓ gives the risk ratios for other variables for all-cause mortality. The only other significant variable was the cause of LV dysfunction. Therefore, we re-examined the predictive value of PNE for all-cause mortality as a subset of patients with ischemic LV dysfunction using the same Cox proportional hazards model. Even in this group, PNE remained the most important predictor for all-cause mortality (risk ratio, 2.91; 95% CI, 1.36 to 6.26; P=.006; Table 3⇓). It is important to note that the significant association between PNE levels and all-cause mortality or cardiovascular mortality was demonstrable, despite adjustment for the effect of treatment with enalapril and other clinical variables.
Hospitalization for Heart Failure Symptoms
Altogether, 65 participants (26%) with PNE levels above the median died or were hospitalized for new or worsening heart failure compared with only 26 (10.2%) of those with PNE levels below the median (Fig 1C⇑; risk ratio, 2.55; 95% CI, 1.6 to 4.1; P<.0001; Fig 2⇑). There were 36 participants (14%) hospitalized for CHF in participants with PNE levels above the median compared with 15 (5.9%) of those with PNE levels below the median (Fig 1D⇑; risk ratio, 2.4; 95% CI, 1.3 to 4.47; P=.005; Fig 2⇑). The median lengths of time to the first hospitalization for heart failure were 440 and 538 days for those with PNE levels above and below the median, respectively.
Development of Heart Failure
In those with PNE levels above the median, 71 participants (27.8%) developed or complained of worsening heart failure symptoms. In contrast, only 38 participants (15%) with PNE levels below the median developed heart failure symptoms (Fig 1E⇑; risk ratio, 1.88; 95% CI, 1.26 to 2.81; P=.002; Fig 2⇑). The median lengths of time required for development of heart failure symptoms in participants with PNE levels above and below the median were 844 and 868 days, respectively. Regardless of the definition used for development of heart failure,16 a significant increase in the development of heart failure was observed in participants with PNE levels above the median.
Development of Ischemic Events
Of the participants with PNE levels above the median, 72 (28.2%) experienced myocardial infarction or hospitalization for unstable angina during follow-up. In contrast, only 41 participants (16.1%) in the group with PNE levels below the median experienced a similar event (Fig 1F⇑; risk ratio, 1.92; 95% CI, 1.29 to 2.83; P=.001; Fig 2⇑ and Table 4⇓). The median lengths of time required for development of any ischemic event in those with PNE levels above and below the median were 863 and 889 days, respectively. Next we examined the relationship between increased PNE levels and myocardial infarction or hospitalization for unstable angina separately. There were 32 myocardial infarctions (12.6%) in those participants with PNE above the median compared with 12 (4.7%) in those with PNE below the median. The risk ratio for development of myocardial infarction was 2.59 (95% CI, 1.33 to 5.01; P=.005). In contrast, the risk ratio for hospitalization for unstable angina was 1.52 (95% CI, 0.98 to 2.37; P=.06). The median follow-up times for development of myocardial infarction in patients with PNE above and below the median were 865 and 892 days, respectively. Table 4⇓ lists the risk ratios for other variables for development of an ischemic event. Thus, other than PNE, the cause of LV dysfunction was the other significant variable that predicted the subsequent development of an ischemic event in this patient population. Therefore, we re-examined the predictive value of PNE for ischemic events only in the subset of patients with ischemic LV dysfunction using the same Cox proportional hazards model. Even in this subgroup, PNE was the most significant predictor for development of an ischemic event (risk ratio, 1.86; 95% CI, 1.23 to 2.8; P=.003) and myocardial infarction (risk ratio, 2.27; 95% CI, 1.15 to 4.45; P=.019). Thus, increased PNE levels appear to be highly predictive for subsequent development of an ischemic event, particularly myocardial infarction in this population of patients with asymptomatic LV dysfunction.
PRA and Clinical Events
The median PRA level before randomization was 0.6 ng·mL−1·h−1 in 511 participants with an interquartile range of 0.3 to 1.3 ng·mL−1·h−1 (1.2±2.0 ng·mL−1·h−1). During follow-up, there were 34 deaths in those with PRA activity above the median (n=276; 12.3%) and 22 deaths in those with PRA values below the median (n=235; 9.4%). This corresponds to a risk ratio of 1.51 with a 95% CI of 0.87 to 2.63 (P=.14). This difference failed to reach statistical significance (Table 5⇓). Similarly, the frequency of other clinical end points, including the development of unstable angina or myocardial infarction, was higher in participants with prerandomization PRA values above the median but failed to reach statistical significance (Table 2⇑). As with ANP, we examined whether the predictive usefulness of PNE levels was independent of the effect of PRA. As with ANP, the adjustment for PRA in the Cox proportional hazards model did not alter the significance of prerandomization PNE levels on the development of clinical events.
ANP Levels and Clinical Events
Median ANP level before randomization was 95 pg/mL with an interquartile range of 58 to 133 pg/mL in 241 participants (108±67 pg/mL). During follow-up, there were 18 deaths in those participants with ANP levels greater than the median (n=121; 14.9%) and 16 deaths in those with ANP levels below the median (n=120; 13.3%). This corresponds to a risk ratio of 0.8 with a 95% CI of 0.38 to 1.65 (P=.54). Similarly, the development of other clinical end points was not influenced by the differences in prerandomization ANP levels (Table 5⇑). Because a previous study had indicated that in participants with symptomatic CHF, ANP was predictive of mortality,8 we re-examined the predictive usefulness of PNE levels after adjusting for the prerandomization ANP levels in the Cox proportional hazards model. Despite this adjustment, PNE levels continued to be highly predictive for all-cause mortality or development of clinical events, including ischemic events.
Prognostic Importance of PNE
This study demonstrates the importance of initial PNE levels as a powerful prognostic predictor for mortality and morbidity in patients with asymptomatic LV dysfunction. To the best of our knowledge, there are no previous reports related to this finding. These observations extend the previous observations that PNE levels were a predictor of clinical events in participants with symptomatic CHF.17 18 PNE level emerged as the most important predictor of mortality, despite adjustment for possible differences in the LV ejection fraction, NYHA class, age, sex, cause of heart failure, treatment assignment to placebo or enalapril, or prerandomization ANP or PRA levels. We also found that increased PNE levels predicted an increased risk for hospitalization for CHF-related symptoms. Thus, elevated PNE concentrations above the median of 393 pg/mL in patients with asymptomatic LV dysfunction precede and predict subsequent progression of the disease and development of clinical outcomes. This median for PNE is ≈75 pg/mL higher than the median PNE levels (median, 317 pg/mL; interquartile range, 242 to 450 pg/mL) measured in a matching cohort of 56 normal control subjects in the SOLVD trial.9 A significant variation in PNE concentrations among CHF patients may make it difficult to use this hormone for risk stratification in a given patient. However, we have shown that in any given patient with CHF who is on stable medications, PNE levels remained relatively stable over a 3-week period.19 20 Thus, measurement of PNE may be a possible marker for assessment of disease progression in patients with LV dysfunction.
PNE levels do not directly reflect the sympathetic release from nerve endings but are influenced by alterations in neuronal uptake, clearance, and metabolism of norepinephrine released from sympathetic nerve endings. However, microneuroangiographic studies in participants with CHF have established a correlation between sympathetic activity and increased PNE levels.21 Therefore, the increased PNE levels may indicate early activation of the sympathetic nervous system in these participants before the development of congestive symptoms. This elevation in PNE levels may contribute to the progression of the disease syndrome by impairing physiological vasodilation,22 adversely affecting cardiac loading through deleterious effects on myocardial function and structure,23 and inciting development of ventricular arrhythmias.24 Data from two recent trials indicate that β-blocker therapy reduces worsening heart failure in patients with heart failure.25 26 Together with our data, it appears that PNE elevations are casually involved in the progression of LV dysfunction to overt heart failure and further clinical deterioration.
The observation that increased PNE levels also are associated with development of myocardial infarction in patients with asymptomatic LV dysfunction has not been previously demonstrated. The majority of our SOLVD participants had ischemic heart disease as the cause of heart failure. Therefore, it may be argued that those patients with histories of myocardial infarction or other ischemic events who are at a higher risk for developing subsequent ischemic events may have contributed to the association between increased PNE levels and the development of ischemic end points. However, when we examined the cohort of patients with LV dysfunction caused by ischemic heart disease, the increase in PNE levels was still strongly associated with a subsequent likelihood of development of a myocardial infarction. This suggests that elevated PNE levels in this type of patient may adversely contribute to the development of ischemic syndromes. Although the mechanisms by which an increase in catecholamines may cause these events are not completely understood, promotion of platelet aggregation and platelet thrombi formation,24 increased myocardial oxygen demand owing to an increase in heart rate, and predisposition to the development of ventricular arrhythmias are some of the possible explanations. Our findings also are supported by trials of short-term and long-term β-blocker therapy after acute myocardial infarction, indicating a reduction in recurrent myocardial infarction.
Prognostic Significance of Other Neurohormones
Previous studies have indicated that symptomatic participants with CHF with elevated ANP levels have increased mortality rates.8 18 In this SOLVD subset of patients with asymptomatic LV dysfunction, however, the modest increase in ANP level was not associated with increased risk for death or development of other clinical events. This lack of correlation may be due to the smaller sample size or less atrial stretching.27 28 In fact, a significant increase in plasma ANP level may occur only when patients become symptomatic with retention of salt and water.6 Therefore, in the absence of symptoms of heart failure, a marked increase in atrial pressures is unlikely, and circulating levels of ANP may be only mildly elevated, which may not be associated with the development of adverse clinical events. A final possibility is that N-terminal ANP and not C-terminal ANP (as measured in the present study) may be a more suitable neurohormonal marker in patients with asymptomatic LV dysfunction.29
Other studies in patients with symptomatic CHF have demonstrated a correlation between increased PRA and mortality.8 24 Although in our cohort of participants an increase in PRA was associated with an increase in risk for development of clinical events, it was not statistically significant. Previous studies have demonstrated the role of diuretics in increasing PRA in CHF patients.9 11 Francis et al9 showed that in patients with asymptomatic LV dysfunction, PRA was normal in the absence of diuretic treatment. In this larger cohort of participants with asymptomatic LV dysfunction, PRA was similarly normal (0.6 ng·mL−1·h−1). Thus, our participants at prerandomization had little or no activation of the renin-angiotensin system, and in the absence of significant stimulation of the renin-angiotensin axis, PRA appears unlikely to be associated with the development of clinical events. It is possible that other hormones, such as angiotensin II, aldosterone, or endothelin, may be just as useful as PNE in predicting the development of clinical events. However, this was not measured in the SOLVD trial, which is a limitation.
Recently the SAVE study investigators examined the prognostic significance of neurohormonal activation in patients after acute myocardial infarction.30 In this study, by multivariate analysis, PNE was only weakly predictive for subsequent development of severe heart failure or combined cardiovascular end points. This most likely is due to the differences in the population examined (after myocardial infarction and with higher ejection fractions) in the SAVE trial compared with the Prevention Trial participants with asymptomatic LV dysfunction. The mean ejection fraction in SAVE was 0.31 compared with 0.28 in SOLVD. This is supported by the fact that the SAVE trial patients in the placebo group had a PNE level of 289±171 pg/mL, which was significantly lower than the median or mean PNE levels in our study population. Thus, elevation of PNE levels appears to be a characteristic in patients with asymptomatic and symptomatic LV dysfunction and may be a reflection of the decrease in LV ejection fraction. Similarly, lower levels of PRA (2.9±3.3 ng·mL−1·h−1) and ANP (70±74 pg/mL) were reported for the SAVE trial patients. Thus, the lack of correlation in our study between PRA or ANP levels and adverse clinical outcomes, unlike in the SAVE trial, also is probably due to the differences in patient population examined.
Conclusions and Implications
There is a stepwise increase in PNE concentrations as subjects change from normal to having asymptomatic LV dysfunction and finally symptomatic heart failure. Elevated levels of PNE but not ANP or PRA predicted cardiovascular mortality and morbidity in patients with asymptomatic LV dysfunction. Although PNE levels are less elevated in this cohort of patients compared with patients with symptoms, the importance of this study is that even modest elevations in PNE levels had significant prognostic importance. While speculative, these data suggest that modulating the release or effect of PNE may lead to improved prognosis and/or reduced mortality and morbidity in this population. This hypothesis should be evaluated in large, prospective, randomized studies.
Selected Abbreviations and Acronyms
|ANP||=||atrial natriuretic peptide|
|CHF||=||congestive heart failure|
|NYHA||=||New York Heart Association|
|PRA||=||plasma renin activity|
|SAVE||=||Survival and Ventricular Enlargement|
|SOLVD||=||Studies of Left Ventricular Dysfunction|
This study was supported by contracts from Studies of Left Ventricular Dysfunction, Clinical Trials Branch, NHLBI, NIH, Bethesda, Md.
Reprint requests to C.R. Benedict, MD, Department of Internal Medicine, Division of Cardiology, University of Texas Medical School, 6431 Fannin MSB 6.039, Houston, TX 77030.
The guest editor for this article was Thomas W. Smith, MD, Brigham and Women's Hospital, Boston, Mass.
- Received June 23, 1995.
- Revision received February 16, 1996.
- Accepted February 20, 1996.
- Copyright © 1996 by American Heart Association
Cody RJ, Atlas SA, Laragh JH, Kubo SH, Covit AB, Ryman KS, Shaknovich A, Pondolfino K, Clark M, Camargo MJF, Scarborough RM, Lewicki JA. Atrial natriuretic factor in normal subjects and heart failure patients: plasma levels and renal hormonal and hemodynamic responses to peptide infusion. J Clin Invest. 1986;78:1362-1374.
Curtiss C, Cohn JN, Vrobel T, Franciosa JA. Role of renin-angiotensin system in systemic vasoconstriction of chronic congestive heart failure. Circulation. 1978;58:763-770.
Sigurdsson A, Amtorp O, Gundersen T, Nilsson B, Jouko R, Swedberg K, for the Ramipril Trial Study Group. Neurohormonal activation in patients with mild or moderately severe congestive heart failure and effect of ramipril. Br Heart J. 1994;72:422-427.
Lee WH, Packer M. Prognostic importance of serum sodium concentration and its modification by converting-enzyme inhibition in patients with severe chronic heart failure. Circulation. 1986;73:257-262.
Francis GS, Benedict CR, Johnstone DE, Kirlin P, Nicklas J, Liang, CS, Kubo SH, Rudin-Toretsky E, Yusuf S, for SOLVD Investigators. Comparison of neuroendocrine activation in patients with left ventricular dysfunction with and without congestive heart failure. Circulation. 1990;82:1724-1729.
Benedict CR, Johnstone DE, Weiner DH, Bourassa MG, Bittner V, Kay R, Kirlin P, Greenberg B, Kohn RM, Nicklas JM, McIntyre K, Quinones MA, Yusuf S, for the SOLVD Investigators. Relation of neurohumoral activation to clinical variables and degree of ventricular dysfunction: a report from the registry of studies of left ventricular dysfunction. J Am Coll Cardiol. 1994;23:1410-1420.
Hussain MN, Benedict CR. Radioenzymatic microassay for simultaneous estimations of dopamine, norepinephrine and epinephrine in plasma, urine and tissues. Clin Chem. 1985;31:1861-1864.
Sealey J, Laragh JH. How to do a plasma renin assay. In: Laragh JG, ed. Topics in Hypertension. New York, NY: York Medical Books; 1980:244-256.
Bodola F, Benedict CR. Rapid simplified radioimmunoassay of arginine vasopressin and atrial natriuretic peptide in plasma. Clin Chem. 1988;34:970-973.
Fleiss JL, Bigger JT Jr, McDermott M. Nonfatal myocardial infarction is, by itself, an inappropriate end point in clinical trials in cardiology. Circulation. 1990;81:684-685.
Cohn J, Levine TB, Olivari MT, Garberg V, Lura D, Francis GS, Simon AB, Rector TS. Plasma norepinephrine as a guide to prognosis in patients with chronic congestive heart failure. N Engl J Med. 1984;311:819-823.
Swedberg K, Eneroth P, Kjekshus J, Wilhelmsen L, for the CONSENSUS Trial Study Group. Hormones regulating cardiovascular function in patients with severe congestive heart failure and their relation to mortality. Circulation. 1990;82:1730-1736.
Benedict CR, Francis G, Shelton BJ, Johnstone D, Kubo S, Kirlin PC, Nicklas J, Liang C-S, Konstam MA, Greenberg B, Yusuf S, for the SOLVD Investigators. Effect of long-term enalapril therapy on neurohormones in patients with left ventricular dysfunction. Am J Cardiol. 1995;75:1151-1157.
Leimbach WN, Wallin G, Victor RG, Aylward PE, Sundlof G, Mark AL. Direct evidence from intraneuronal recordings for increased central sympathetic outflow in patients with heart failure. Circulation. 1986;73:913-919.
Francis GS, McDonald KM, Cohn JN. Neurohumoral activation in preclinical heart failure: remolding and the potential for intervention. Circulation. 1993;87(suppl IV):IV-90-IV-96.
Benedict CR. Neurohumoral aspects of heart failure. Cardiol Clin North Am. 1994;12:9-23.
Waagstein F, Bristow MR, Swedberg K, Caprerini F, Folwer MB, Silver MA, Gilbert EM, Johnson MR, Goss FG, Hjalmarson A, for the Metoprolol in Dilated Cardiomyopathy (MDC) Trial Study Group. Beneficial effects of metoprolol in idiopathic dilated cardiomyopathy. Lancet. 1993;342:1441-1446.
CIBIS Investigators. A randomization trial of β-blockade in heart failure: the Cardiac Insufficiency Bisoprolol Study (CIBIS). Circulation. 1994;90:1765-1773.
Bates ER, Shenker Y, Grekin RJ. The relationship between plasma levels of immunoreactive atrial natriuretic hormone and hemodynamic function in man. Circulation. 1986;73:1155-1161.
Rouleau JL, Packer M, Moye L, De Champlain J, Bichet D, Klein M, Rouleau JR, Sussex B, Arnold JM, Sestier F, Parker JO, McEwan P, Bernstein V, Cuddy TE, Lamas G, Gottlieb SS, McCans J, Nadeau C, Delage F, Wun CC, Pfeffer MA. Prognostic value of neurohormonal activation in patients with an acute myocardial infarction: effect of captopril. J Am Coll Cardiol. 1994;24:583-591.