Valve Disease METHODS

Journal of the American College of Cardiology Vol. 36, No. 1, 2000 2000 by the American College of Cardiology ISSN 0735-1097/00/$20.00 Published by Elsevier Science Inc. PII S0735-1097(00)00721-X Valve Disease Elevated Circulating Levels of Serum Tumor Necrosis Factor-alha in Patients With Hemodynamically Significant Pressure and Volume Overload Samir R. Kaadia, MD, Khalid Yakoob, MD, Simone Nader, MD, James D. Thomas, MD, Douglas L. Mann, MD,* Brian P. Griffin, MD Cleveland, Ohio and Houston, Texas OBJECTIVES BACKGROUND METHODS RESULTS CONCLUSIONS We sought to determine whether serum tumor necrosis factor-alha (TNF-alha) levels are elevated in atients with hemodynamically significant ressure and volume overload. It has been reviously shown that TNF-alha messenger ribonucleic acid (mrna) and rotein are raidly exressed in the hearts of animal models subjected to abrut hemodynamic overloading. The clinical significance of these exerimental findings has not been tested in athohysiologically relevant clinical models in human subjects. We rosectively measured serum TNF-alha levels and serum TNF recetor 1 and 2 levels in 21 atients with severe aortic stenosis (AS), in 26 atients with 3 to 4 mitral regurgitation (MR) and in normal age- and gender-matched control subjects. Patients with AS and MR were either in New York Heart Association (NYHA) functional class I or II and had no significant coronary disease. We comared the cytokine levels among the grous using analysis of variance. We related cytokine levels to the severity of AS using simle regression analysis. Serum TNF-alha levels in atients with AS (2.1 1.6 g/ml, n 21) and MR (1.3 0.7 g/ml, n 26) were significantly higher than those in the control subjects (0.7 0.2 g/ml, n 28). Serum TNF recetor 1 and 2 levels were also higher in atients with AS and MR than in control subjects. Cytokine levels were higher in atients in NYHA class II than in those in class I. In atients with a normal ejection fraction ( 50%, n 16), there was a mild ositive correlation (r 0.56, 0.025) between serum TNF-alha levels and the mean gradient across the aortic valve. This study demonstrates that serum TNF-alha is elevated in atients with chronic hemodynamic overloading and early cardiac decomensation. Furthermore, these findings suggest not only that eriheral TNF-alha levels correlate with the severity of the hemodynamic ressure overload, but also that eriheral TNF-alha and TNF recetor levels increase in direct relation to deteriorating NYHA functional class. (J Am Coll Cardiol 2000; 36:208 12) 2000 by the American College of Cardiology There is increasing evidence that the rogressive elaboration of tumor necrosis factor-alha (TNF-alha) may lay a role in the cardiac decomensation that ensues in a variety of cardiac disease states (1 5). Although TNF-alha was initially described as a myocardial deressant substance in atients with setic shock, elevation of this cytokine has subsequently been demonstrated in a variety of cardiac disease states, including those of atients with symtomatic heart failure (6,7). When overexressed, TNF-alha can exert direct toxic effects on the heart and the eriheral circulation, and thus may otentially exlain some of the structural and functional changes that are observed in cardiac decomensation. Indeed, overexression of TNFalha has been shown to lead to left ventricular (LV) dysfunction and LV remodeling in animal models with From the Cleveland Clinic Foundation, Cleveland, Ohio; and *Winters Center for Heart Failure Research, Baylor College of Medicine and Veterans Affairs Medical Center, Houston, Texas. Manuscrit received July 22, 1999; revised manuscrit received January 20, 2000, acceted March 27, 2000. long-term infusion of TNF-alha, as well as in transgenic animals with targeted cardiac overexression (1,2,8). In the natural history of valvular heart disease, there is a hase of comensatory hyertrohy followed by rogressive cardiac decomensation. However, the exact mechanisms that are resonsible for the develoment of cardiac decomensation after hemodynamic overloading remain unknown. Relevant to this discussion is the exerimental observation that hemodynamic overloading and myocardial stretch are sufficient to rovoke de novo TNF-alha messenger ribonucleic acid (mrna) and rotein exression in the heart (9,10). However, the clinical significance of these early exerimental studies is not known. Accordingly, the urose of this study was to rosectively study TNF-alha and TNF recetor levels in atients with clinically significant aortic stenosis (AS) or mitral regurgitation (MR). METHODS Patient grou. We rosectively enrolled atients with an echocardiograhic diagnosis of either significant AS (aortic

JACC Vol. 36, No. 1, 2000 July 2000:208 12 Kaadia et al. TNF-alha and Hemodynamic Overloading 209 Abbreviations and Acronyms AS aortic stenosis AVA aortic valve area CI confidence interval EF ejection fraction LV left ventricular mrna messenger ribonucleic acid MR mitral regurgitation NYHA New York Heart Association OR odds ratio TNF-alha tumor necrosis factor-alha valve area [AVA] 1.0 cm 2 ) or significant MR (3 to 4 ), who had mild symtoms (New York Heart Association [NYHA] functional class I or II). The severity of MR was determined semiquantitatively by using the regurgitant jet length, width and direction on color Doler echocardiograhy (11). Every atient had cardiac catheterization within six months of enrollment, and atients with coronary artery stenosis 50% were excluded. Patients with diseases associated with raised serum TNF-alha, such as recent endocarditis (within six months), active malignancies, infection, connective tissue disorders and fever, were excluded. Patients with comlex multivalvular lesions were excluded. For the control grou, middle-aged (55 8 years old, n 28, including 16 [57%] men) normal subjects with no significant illness were used. Serum cytokine measurements. Blood was collected in a nonhearinized glass tube from a forearm vein. Blood was stored on ice immediately after collection, and within 15 min serum was searated by standard centrifugation at 1,500 rm for 10 min. Serum was immediately searated and stored at 70 C until the assay. Cytokine measurements were erformed in a core laboratory by scientists who had no knowledge of the clinical information. Commercially available enzyme-linked immunosorbent assay kits (Quantikine HS, R&D Systems) were used for measurement of TNF-alha and TNF recetors 1 and 2, exactly as described reviously (4). Clinical data collection. The clinical information was collected rosectively before the assessment of eriheral levels of cytokine and cytokine recetors. The atients functional class was determined on the basis of the clinical history, as well as from chart review. Echocardiograhy was erformed in the research echocardiograhy laboratory at the Cleveland Clinic. M-mode, two-dimensional and Doler echocardiograhic measurements were made on-line by indeendent observers who had no knowledge of the clinical history. We took end-diastolic measurements of osterior wall and intraventricular setal thickness and end-systolic and end-diastolic measurement of LV diameter, left atrial diameter, ejection fraction (EF) by Simson s rule and LV mass measurement by the Penn Convention (12,13). Continuous wave Doler echocardiograhy was used to measure the severity of AS. Peak and mean gradients, as well as Table 1. Clinical Characteristics Patient Characteristics AVA, using the continuity equation, were calculated for each atient. As reviously described, the severity of MR was assessed semiquantitatively by color flow maing (14) and by analysis of ulmonary venous flow (15 17). Only atients with significant MR ( 3 ) were included in the study. Statistical analysis. Data are exressed as the mean value SD. Continuous variables were comared using the unaired, two-tailed Student t test. When comaring continuous variables among multile grous, analysis of variance, with ost hoc analysis using the Newman-Keuls test, was emloyed. Simle linear regression analysis was erformed to study the relation between continuous variables. Multile regression analysis was erformed to determine whether the resence of AS and MR was associated with higher TNF-alha levels indeendent of age and gender. RESULTS (n 21) (n 26) Age (yrs) 71 12 56 13 0.0004 Male gender 13 (62%) 17 (65%) NS Hyertension 13 (68%) 21 (81%) NS Diabetes mellitus 3 (17%) 1 (4%) NS Smoking 10 (56%) 7 (27%) NS History of AF 3 (15%) 8 (30%) NS NYHA class I 6 (29%) 12 (46%) NS SBP/DBP (mm Hg) 133 15/75 8 119 31/73 20 NS Serum sodium 139 3 141 2 NS (meq/liter) Serum BUN (mg/dl) 18 7 16 4 NS Serum creatinine (mg/dl) 1.0 0.2 0.9 0.2 NS WBC count (10 3 /mm 3 ) 7 2 6 1 NS Medications ACE inhibitors 3 (14%) 7 (27%) NS Digoxin 6 (28%) 6 (23%) NS Diuretics 2 (10%) 8 (31%) NS Data are resented as the mean value SD or number (%) of atients. ACE angiotensin-converting enzyme; AF atrial fibrillation; BUN blood urea nitrogen; DBP diastolic blood ressure; NS not significant; NYHA New York Heart Association; SBP systolic blood ressure; WBC white blood cell. Patient grou. We studied 47 mildly symtomatic atients with either AS or MR. Of these 47 atients, 21 had severe AS and 26 had severe MR. The clinical characteristics of these atients are listed in Table 1. Patients with AS were significantly older as comared with atients with MR. The use of angiotensin-converting enzyme inhibitors, digoxin and diuretics for each grou are summarized in Table 1. The echocardiograhic characteristics are listed in Table 2. The EF was reduced to 50% in five atients with AS and in one atient with MR. Circulating levels of TNF-alha and TNF recetors. Serum TNF-alha levels in atients with AS and MR were significantly higher than those levels in the control grou (Fig. 1). Patients with AS had significantly higher ( 0.05) TNF-alha levels than did atients with MR. Even

210 Kaadia et al. JACC Vol. 36, No. 1, 2000 TNF-alha and Hemodynamic Overloading July 2000:208 12 Table 2. Echocardiograhic Characteristics Echocardiograhic Characteristics (n 21) (n 26) Ejection fraction (%) 54 14 63 8 LV end-systolic diameter (mm) 28 9 35 5 LV end-diastolic diameter (mm) 47 9 62 7 Setal thickness (mm) 14 3 11 3 Posterior wall thickness (mm) 14 3 10 2 LA diameter (mm) 42 9 50 12 LV mass index (g/m) 186 59 199 68 valve area (cm 2 ) 0.65 0.13 Peak/mean gradient (mm Hg) 96 37/57 21 Data are resented as the mean value SD. LA left atrial; LV left ventricular. when atients in NYHA class I were analyzed, the TNFalha levels in atients with AS were higher than those in atients with MR (Table 3). In a multile regression analysis, the resence of AS (odds ratio [OR] 1.4, 95% confidence interval [CI] 1.13 to 1.85) and MR (OR 1.2, 95% CI 1.03 to 1.55) redicted higher TNF-alha levels, even after adjusting for the age differences. Interestingly, in atients with AS with reserved ventricular function (EF 50%, n 16), there was a modest ositive correlation (r 0.56, 0.025) between eriheral levels of TNF-alha level and the mean gradient across the aortic valve (Fig. 2). The circulating levels of soluble TNF recetors were also significantly higher in atients with AS than in the control subjects ( 0.001 for both) (Fig. 3A). However, in atients with MR, TNF recetors 1 and 2 were only minimally elevated and were not statistically different from those in control subjects ( 0.11 and 0.18, resectively) (Fig. 3B). For atients with AS and MR, TNF recetors 1 and 2 were significantly higher in atients in NYHA class II than in atients in NYHA class I (Table 4). Finally, there was a modest correlation between the serum level of TNF recetors 1 and 2 and the mean gradient across Table 3. Serum Tumor Necrosis Factor-Alha Levels in Patients With and, According to New York Heart Association Functional Class the aortic valve (r 0.44, 0.047 and r 0.43, 0.05, resectively). DISCUSSION All atients 2.1 1.6 1.3 0.7 0.03 Class I 1.7 0.7 0.9 0.5 0.02 Class II 2.2 1.9 1.6 0.7 0.23 Data are resented as the mean value SD in g/ml. This study shows that circulating levels of TNF-alha are elevated in atients with clinically significant hemodynamic ressure and volume overload. Moreover, it suggests not only that the eriheral TNF-alha levels correlate with the severity of the hemodynamic ressure overload (Fig. 2), but also that the eriheral TNF-alha and TNF recetor levels increase in direct relation to deteriorating NYHA functional classification (Tables 3 and 4). One of the more interesting, albeit unexected, findings of this study was that atients with chronic ressure overload had higher serum TNF-alha and TNF recetor (recetors 1 and 2) levels than did atients with chronic volume overload, thus suggesting that ressure overload may serve as a greater stimulus for cytokine roduction, as comared with volume overload. However, the recise reasons for these differences in cytokines and cytokine recetors in atients with ressure and volume overload are not known at this time. Circulating levels of TNF-alha and TNF recetors have consistently been shown to be elevated in atients with heart failure; moreover, the eriheral levels of these roteins have been shown to correlate with the atients NYHA Figure 1. Serum TNF-alha levels in normal subjects (0.7 0.2 g/ml, n 28), in atients with AS (2.1 1.6 g/ml, n 21) and in atients with MR (1.3 0.7 g/ml, n 26). Analysis of variance showed significant differences between the grous, and a ost hoc comarison demonstrated significant differences between all three grous. Serum TNF-alha levels of atients with NYHA class I symtoms and control subjects are shown by solid circles, and those levels of atients with class II symtoms are shown by oen circles. Figure 2. Correlation between the mean gradient across the aortic valve and serum TNF-alha level in atients with AS and a normal ( 50%) LV ejection fraction (n 16). The central line reresents the regression, and the other lines show the 95% confidence levels for the regression.

JACC Vol. 36, No. 1, 2000 July 2000:208 12 Kaadia et al. TNF-alha and Hemodynamic Overloading 211 Table 4. Serum Levels of Tumor Necrosis Factor Recetors in Patients With and, According to New York Heart Association Functional Class TNF-Recetor 1 (g/ml) All atients 1,614 884 1,060 270 0.004 Class I 1,302 330 897 209 0.005 Class II 1,739 1,010 1,199 241 0.06 TNF-Recetor 2 (g/ml) All atients 4,559 2,389 3,201 838 0.010 Class I 3,682 1,038 2,645 743 0.03 Class II 4,910 2,702 3,680 593 0.11 Data are resented as the mean value SD. TNF tumor necrosis factor. Figure 3. A, Soluble TNF recetor 1 levels in normal subjects (841 361 g/ml, n 32), in atients with AS (1,614 884 g/ml, n 21) and in atients with MR (1,060 270 g/ml, n 26). Soluble TNF recetor 1 levels of atients with NYHA class I symtoms and control subjects are shown by solid squares, and those levels of atients with class II symtoms are shown by oen squares. B, Soluble TNF recetor 2 levels in normal subjects (2,694 884 g/ml, n 32), in atients with AS (4,559 2,389 g/ml, n 21) and in atients with MR (3,202 838 g/ml, n 26). Soluble TNF recetor 2 levels of atients with NYHA class I symtoms and control subjects are shown by solid triangles, and those levels of atients with class II symtoms are shown by oen triangles. functional class (4,18). However, the site and source of cytokine roduction in heart failure remain unknown. Relevant to this discussion is the observation that both TNFalha rotein and mrna are exressed in the hearts of atients with advanced heart failure, whereas neither TNFalha rotein nor mrna are exressed in the hearts of normal subjects (5). Although the mechanism for the increase in TNF-alha mrna and rotein exression is not known, a revious study has shown that TNF-alha mrna and rotein are raidly exressed in an animal model of acute ressure overload (9). Accordingly, the ositive correlation between eriheral TNF-alha levels and the mean gradient across the aortic valve in atients with AS suggests not only that chronic ressure overload is a relevant stimulus for TNF-alha exression, but also that the heart is likely to be an imortant source for TNF-alha roduction. Patients with abnormal LV systolic function were excluded from this analysis, as the ressure gradient in these atients does not accurately reflect the severity of AS and ressure overload. Although we observed that there was a significant increase in serum TNF-alha levels in direct relation to deteriorating NYHA class, the results of the resent study do not distinguish whether elevated TNF-alha levels are a cause or effect of increasing cardiac decomensation. Indeed, given the biologic comlexity of hemodynamic overloading, it would be naive to suggest that any one molecule was resonsible for cardiac decomensation. This statement notwithstanding, it bears emhasis that overexression of TNF-alha in the heart has been shown to lead to rogressive LV remodeling, as well as rogressive LV dysfunction in exerimental models (1 3). Moreover, elevated eriheral TNF-alha levels might be exected to lead to inaroriate vasodilation by increasing endothelial nitric oxide roduction, which may have overt deleterious consequences in atients with AS. Thus, it is ossible that TNF-alha may be one of a variety of different biologically active molecules that contribute to the rogressive cardiac decomensation that occurs in sustained hemodynamic overloading. Aart from the otential athohysiologic relevance of the resent findings, the results of this study suggest that rogressively rising TNF-alha levels may serve as a otentially useful biochemical marker for rogressive cardiac decomensation in atients with valvular heart disease. Although it is unlikely that elevated levels of TNF-alha will ever relace the more traditional indexes that have been used in the timing of surgery for atients with symtomatic valvular heart disease, it is ossible that rogressive increases in TNF-alha levels may rove to be useful in the timing of valve relacement in atients with asymtomatic valvular heart disease, in whom a subset of atients exerience untoward outcomes (19 23). One of the major study limitations includes the unavailability of ostoerative TNF-alha levels to verify whether elevated cytokine levels became normalized after surgical correction of ressure or volume overload. Immediate ostoerative levels may not be accurate, as the oeration itself

212 Kaadia et al. JACC Vol. 36, No. 1, 2000 TNF-alha and Hemodynamic Overloading July 2000:208 12 can lead to elevated cytokine levels. Referring hysicians monitor their own atients, which makes it logistically difficult to collect ostoerative cytokine levels. Further, the atients with AS are older than atients with MR and the normal control subjects. Although these differences in age between the atient grous does not comletely account for the elevated cytokine levels, the imact of aging on cytokine roduction remains unknown. Conclusions. This clinical study directly suorts the hyothesis that hemodynamic overloading is a sufficient stimulus for TNF-alha roduction in the adult human heart, which is consistent with our revious exerimental observations. Moreover, this study suggests that eriheral TNFalha and TNF recetor levels increase in relation to the atient s worsening NYHA functional class. Given the mechanical burden that is imosed by sustained hemodynamic overloading, it is unlikely that modulating TNFalha exression will favorably alter the natural history of valvular heart disease. Nonetheless, it is ossible that anticytokine strategies might be emloyed as an adjunctive theray before oeration in high risk atients with deressed ventricular erformance, in whom sustained ventricular dysfunction may revent immediate ostoerative recovery. Moreover, this study raises the intriguing ossibility that serial measurements of TNF-alha in atients with asymtomatic valvular heart disease may hel to identify atients in the earliest stages of decomensation, who are not diagnosed with conventional noninvasive testing. Further studies will be necessary to address these interesting and imortant ossibilities. Acknowledgments The authors acknowledge the technical assistance of Lisa Cardon, BS, and Dorellyn Lee-Jackson, BS. Rerint requests and corresondence: Dr. Brian P. Griffin, The Cleveland Clinic Foundation, 9500 Euclid Avenue, F15, Cleveland, Ohio 44195. E-mail: Griffib@ccf.org. REFERENCES 1. Bryant D, Becker L, Richardson J, et al. Cardiac failure in transgenic mice with myocardial exression of tumor necrosis factor-alha. Circulation 1998;97:1375 81. 2. Bozkurt B, Kribbs SB, Clubb FJJ, et al. Pathohysiologically relavent concentrations of tumor necrosis factor-alha romote rogressive left ventricular dysfunction and remodeling in rats. Circulation 1998;97: 1382 91. 3. Kubota T, McTiernan CF, Frye CS, et al. Dilated cardiomyoathy in transgenic mice with cardiac-secific overexression of tumor necrosis factor-alha. Circ Res 1997;81:627 35. 4. Torre-Amione G, Kaadia S, Benedict C, Oral H, Young JB, Mann DL. Proinflammatory cytokine levels in atients with deressed left ventricular ejection fraction: a reort from the Studies Of Left Ventricular Dysfunction (SOLVD). J Am Coll Cardiol 1996;27:1201 6. 5. Torre-Amione G, Kaadia S, Lee J, et al. Tumor necrosis factor-alha and tumor necrosis factor recetors in the failing human heart. Circulation 1996;93:704 11. 6. Levine B, Kalman J, Mayer L, Fillit HM, Packer M. Elevated circulating levels of tumor necrosis factor in severe chronic heart failure. N Engl J Med 1990;323:236 41. 7. Matsumori A, Yamada T, Suzuki H, Matoba Y, Sasayama S. Increased circulating cytokines in atients with myocarditis and cardiomyoathy. Br Heart J 1994;72:561 6. 8. Kubota T, McTiernan CF, Frye CS, Demetris AJ, Feldman AM. Cardiac-secific overexression of tumor necrosis factor-alha causes lethal myocarditis in transgenic mice. J Card Failure 1997;3:117 24. 9. Kaadia SR, Oral H, Lee J, Nakano M, Taffet GE, Mann DL. Hemodynamic regulation of tumor necrosis factor-alha gene and rotein exression in adult feline myocardium. Circ Res 1997;81:187 95. 10. Kaadia S, Lee J, Torre-Amione G, Birdsall HH, Ma TS, Mann DL. Tumor necrosis factor-alha gene and rotein exression in adult feline myocardium after endotoxin administration. J Clin Invest 1995;96:1042 52. 11. Miyatake K, Izumi S, Okamoto M, et al. Semiquantitative grading of severity of mitral regurgitation by real-time two-dimensional Doler flow imaging technique. J Am Coll Cardiol 1986;7:82 8. 12. Sahn DJ, DeMaria A, Kisslo J, Weyman A. Recommendations regarding quantitation in M-mode echocardiograhy: results of a survey of echocardiograhic measurements. Circulation 1978;58: 1072 83. 13. Devereux RB, Reichek N. Echocardiograhic determination of left ventricular mass in man: anatomic validation of the method. Circulation 1977;55:613 8. 14. Helmcke F, Nanda NC, Hsiung MC, et al. Color Doler assessment of mitral regurgitation with orthogonal lanes. Circulation 1987;75: 175 83. 15. Pu M, Griffin BP, Vandervoort PM, et al. The value of assessing ulmonary venous flow velocity for redicting severity of mitral regurgitation: a quantitative assessment integrating left ventricular function. J Am Soc Echocardiogr 1999;12:736 43. 16. Klein AL, Obarski TP, Stewart WJ, et al. Transesohageal Doler echocardiograhy of ulmonary venous flow: a new marker of mitral regurgitation severity. J Am Coll Cardiol 1991;18:518 26. 17. Klein AL, Stewart WJ, Bartlett J, et al. Effects of mitral regurgitation on ulmonary venous flow and left atrial ressure: an intraoerative transesohageal echocardiograhic study. J Am Coll Cardiol 1992;20: 1345 52. 18. Ferrari R, Bachetti T, Confortini R, et al. Tumor necrosis factor soluble recetors in atients with various degrees of congestive heart failure. Circulation 1995;92:1479 86. 19. Enriquez-Sarano M, Schaff HV, Orszulak TA, Bailey KR, Tajik AJ, Frye RL. Congestive heart failure after surgical correction of mitral regurgitation: a long-term study. Circulation 1995;92:2496 503. 20. Enriquez-Sarano M, Schaff HV, Frye RL. Early surgery for mitral regurgitation: the advantages of youth (editorial). Circulation 1997;96: 4121 3. 21. Enriquez-Sarano M, Orszulak TA, Schaff HV, Abel MD, Tajik AJ, Frye RL. regurgitation: a new clinical ersective. Mayo Clin Proc 1997;72:1034 43. 22. Starling MR, Kirsh MM, Montgomery DG, Gross MD. Imaired left ventricular contractile function in atients with long-term mitral regurgitation and normal ejection fraction. J Am Coll Cardiol 1993; 22:239 50. 23. Starling MR. Effects of valve surgery on left ventricular contractile function in atients with long-term mitral regurgitation. Circulation 1995;92:811 8.