Original Contribution

Effects of on and Left Ventricular Dysfunction in Male Veterans

Spyros Kokolis, MD, Jonathan D. Marmur, MD, Luther T. Clark, MD, John Kassotis, MD, Rodamanthos Kokolis, MD, Erdal Cavusoglu, MD, Reuven Lapin, PA-C, *Sheldon Breitbart, MD, Jason M. Lazar, MD

ABSTRACT: Background. Heavy consumption is a well- and nonhemorrhagic stroke.3 The coronary protective effects of known cause of dilated and hypertension, but its effects heavy alcohol consumption have not been well studied. Chronic on coronary atherosclerosis are less well understood. The objective of this heavy alcohol consumption is a well-known cause of dilated car- study was to compare coronary anatomy and left ventricular dysfunction diomyopathy. The deleterious effects of heavy alcohol on left ven- in patients with and without alcoholism associated with heavy consump- tricular function and on arrythmogenesis may counterbalance any tion. Methods. We studied 100 consecutive alcoholic male patients pre- 4 senting with chest pain to the Department of Veterans Affairs New York potential coronary benefits. Few studies have addressed both Harbor Healthcare System (VA) between 1994 and 2002. Alcoholism CAD and left ventricular function in patients chronically con- was defined as a history of either chronic alcohol-related pancreatitis or suming large quantities of alcohol. Accordingly, the objective of liver cirrhosis. Patients were compared to age-matched controls (n = 200) this study was to assess the associations between alcoholism, that were known to be nonalcoholic. All patients underwent coronary CAD severity and left ventricular systolic dysfunction. angiography. Results. Baseline demographic characteristics were simi- lar between the two groups. The prevalence of significant coronary Methods artery disease (CAD) (defined as coronary arterial luminal diameter The local institutional review board of the Department of stenosis > 50%) was lower in the alcoholic group than in the control Veterans Affairs New York Harbor Healthcare System approved group (42% vs. 58%; p = 0.013). Among patients with CAD, those the study (VA). The VA internal medicine healthcare database with a history of alcoholism had fewer vessels with stenoses (1.6 ± 0.6 vs. 2.3 ± 0.7; p < 0.001) than the control group, and were more likely to was queried for patients admitted between 1994 and 2002 who have single-vessel CAD (64% vs. 8%; p < 0.05). The alcoholic group also underwent and had a history of chronic had lower mean left ventricular ejection fraction (LVEF) compared to alcoholic pancreatitis or chronic alcoholic liver cirrhosis (diag- the control group (43 ± 13% vs. 49 ± 9%; p < 0.001), and a higher nosed by ICD-9 codes 303.91, 577.1 and 571.2, respectively). prevalence of left ventricular dysfunction (LVEF < 40%; 37% vs. 13%; p < All of these patients had presented to the emergency room 0.05). In the alcoholic group, there was a lower prevalence of CAD in because of chest pain and subsequently had a positive stress test. patients with left ventricular dysfunction as compared to those without The stress test was considered positive if the results of the test left ventricular dysfunction (21% vs. 49%; p = 0.006). Conclusions. In on the chart were documented as “positive”, “equivocal” or a group of male VA patients presenting with chest pain, alcoholism was “nondiagnostic”. Therefore, the patient had this stress test result associated with a lower incidence and a lesser severity of angiographically- defined CAD, but had greater left ventricular dysfunction. There appears followed up with a cardiac angiogram in the VA system. During to be an inverse relationship between CAD and left ventricular function the index admission, all patients underwent in patients with a history of heavy alcohol consumption. and had blood drawn for liver function testing. The patients were then subsequently compared to the first 200 consecutive J INVASIVE CARDIOL 2006;18:304–307 age-matched controls without alcoholism between 1994 and 2002 from the VA internal medicine healthcare database. These Light-to-moderate alcohol consumption has been associated patients were defined as nonalcoholic according to the history 1 with a decreased risk of ischemic cardiac events and stroke. The and physical exams conducted by their outpatient internal med- cardioprotective effects of alcohol have been attributed to favor- icine physicians that were documented in the computers of the able lipid changes, including lower LDL, increased HDL choles- VA internal medicine healthcare. Similar to the alcoholic terol and higher apolipoprotein AI and AII levels, antiplatelet, patients, these nonalcoholic patients were admitted through the 2 and anti-inflammatory effects. Low levels of alcohol consump- emergency room because of chest pain and had a stress test that tion have been proven beneficial in providing a protective effect was documented as “positive”, “equivocal”, or “nondiagnostic”, upon the cerebral circulation. However, in heavy alcohol con- thus prompting the hospital to refer the patient for cardiac sumption, there is an increased predisposition to hemorrhagic angiography within the VA system (Figure 3). The following clinical variables were considered in all patients: From the S.U.N.Y. Downstate Medical Center and the *Department of Veter- ans Affairs New York Harbor Healthcare System, Brooklyn, New York. family history of CAD, active smoking, hypertension, previous The authors report no financial relationships or conflicts of interest regarding the (MI) and diabetes. In both groups of content herein. Manuscript submitted November 29, 2005, provisional acceptance given January patients, the serum laboratory values included: total cholesterol, 3, 2006, manuscript accepted April 3, 2006. triglycerides and high-density lipoproteins (HDL). Low-density Address for correspondence: Jonathan D. Marmur, MD, S.U.N.Y. Downstate lipoproteins (LDL) were determined by the Frederich calcula- Medical Center, Department of , Box 1257, 450 Clarkson Avenue, Brook- lyn, NY 11203. E-mail: [email protected] tion in patients whose triglycerides were < 400 mg/dL.

304 The Journal of Invasive Cardiology Effects of Alcoholism on CAD and LV Dysfunction in Male Veterans

mean ± standard deviation and 70 P = 0.013 P < 0.001

n were compared using the Stu-

o 60 i t dent’s t-test. Categorical vari- 60 c a r ables were expressed as F n D 50 frequencies and percentages and o i A t C c compared using the Fisher’s e f 40 j o

40 E exact test. A p-value of 0.05 was e r c a

l considered significant; all tests n u e l c i

a 30 were two-sided. Clinical, mor- r v t e n

r phological and procedural vari- e

P 20 20 V

t ables that had demonstrated f e

L statistically significant differ-

10 n

a ences among the two groups e

M were included in the stepwise 0 0 Alcoholic Nonalcoholic Alcoholic Nonalcoholic multivariate logistic analysis to determine the effect of alco- Figure 1. (A) (B) The prevalence of coronary artery disease (CAD) and mean left ventricular ejection fraction in alcoholics versus nonalcoholics. holic status as an independent predictor of CAD, left ventric- ular ejection fraction and vari- 100 ous risk factors. A Alcoholic 100 Nonalcoholic B P = 0.48 P < 0.001 80 80 Results A total of 100 patients were SGPT < 40 IU/L SGPT < 40 IU/L identified as having alcohol- e 60 SGPT > 40 IU/L e SGPT > 40 IU/L g g 60 a a related liver cirrhosis or pancre- t t n n e e atitis and had undergone c c r 40 r e e coronary angiography. The P

P 40 mean age was 63 ± 4 years. Ninety-one percent (91%) had 20 20 chronic liver cirrhosis and 9% had a history of alcoholic pan- 0 creatitis. These patients were ≥ 0 EF < 40 EF 40 EF < 40 EF ≥ 40 compared to 200 consecutive Ejection Fraction Ejection Fraction nonalcoholic patients who also Figure 2. The prevalence of SGPT levels and mean left ventricular ejection fraction in alcoholics (A) versus had undergone cardiac catheter- nonalcoholics (B). ization and had a history of chest pain and a positive stress test. Diabetes was more preva- lent in the alcoholic group (59% All coronary angiograms were performed at the VA in Brook- vs. 31%; p < 0.05). Other baseline demographics were similar lyn, New York. CAD was defined visually as an obstructive between both groups (Table 1). However, the prevalence of lesion > 50% stenosis of an epicardial coronary artery. The CAD was lower in the alcoholic group than in the control group branches of major coronary arteries were not assessed in this (42% vs. 58%; p = < 0.05). In patients with CAD, the alcoholic study. The number of stenosed vessels in the major coronary group had fewer mean numbers of stenosed vessels (1.6 ± 0.6 vs. artery distributions — the left main coronary artery, the left 2.3 ± 0.7; p < 0.05), and was more likely to have single-vessel anterior descending artery, the left circumflex artery, and the CAD as compared to the control group (64% vs. 8%; p < 0.05). right coronary artery were visualized to determine if they had an The alcoholic group had a lower mean LVEF compared to obstructive lesion > 50% stenosis to quantify the extent and the control group (43 ± 13 % vs. 49 ± 9 %; p < 0.05), and left severity of CAD in the alcoholic and nonalcoholic patients. ventricular systolic dysfunction was more common (37% vs. LVEF was determined by echocardiography using the Teich- 13%; p < 0.05) than in the nonalcoholic group. In patients holz method. Sensitivities, specificities, positive and negative with left ventricular dysfunction, there was a lower prevalence predictive values of elevated serum SGOT (≥ 40 IU/L) and of CAD in alcoholics as compared to nonalcoholics (38% vs. SGPT (≥ 40 IU/L) values were determined. Left ventricular 63%; p < 0.05). In contrast, in patients with preserved systolic dysfunction was defined as an ejection fraction < 40%.5 function, there was a numerically higher prevalence of CAD Statistical analysis was performed using SPSS software (SPSS, in alcoholics vs. nonalcoholics, although this did not reach sta- Inc., Chicago, Illinois). Continuous variables were expressed as tistical significance (54% vs. 38%; p = NS).

Vol. 18, No. 7, July 2006 305 KOKOLIS, et al.

Table 1. Patient baseline characteristics. Discussion This study showed less prevalent and less extensive CAD in Alcohol Nonalcohol Characteristics p-value alcoholic male patients as compared to the age-matched controls (n = 100) (n = 200) with a history of chest pain and a positive stress test. In addition, Age (years) 63 ± 4 65 ± 3 0.56 left ventricular dysfunction was more prevalent and more severe Hypertension (%) 61 64 0.70 in the alcoholic group (Figure 2). The study extends the observa- Family History (%) 45 41 0.53 tions of epidemiological studies which have shown light to mod- Myocardial Infarction (%) 25 32 0.28 erate alcohol ingestion associated with fewer cardiovascular Diabetes (%) 59 31 < 0.001 events, including MI, cardiovascular mortality, and ischemic Smoking (%) 65 66 0.90 stroke.6,7 These studies evaluated the effect and dose response of Total cholesterol (mg/dL) 174 37 179 32 0.25 ± ± alcohol on a number of clinical endpoints. Proposed mecha- LDL cholesterol (mg/dL) 103 ± 32 106 ± 32 0.53 nisms for decreased cardiovascular events include a favorable HDL cholesterol (mg/dL) 46 ± 11 48 ± 9 0.60 impact in lipids including: increase of HDL, and lowering of Triglycerides (mg/dL) 132 ± 67 134 ± 64 0.81 1 Weight (kg) 79 ± 18 86 ± 19 0.95 LDL by alcohol. There is also an inhibition of platelet adhesion, Body mass index (m2) 24 ± 2 26 ± 4 0.80 increased prostacyclin: thromboxane ratio, increased endogenous Coronary artery disease (%) 42 58 0.013 tissue type plasminogen enhanced anti-oxidant effect, and anti- No. of diseased vessels in inflammatory effects with the exception of a higher prevalence patients with CAD 1.6 ± 0.6 2.3 ± 0.7 < 0.001 of diabetes in the alcoholic group.8 These factors may act collec- LVEF < 40% (%) 37 13 < 0.05 tively to inhibit plaque progression and promote stabilization. Mean LVEF (%) 43 ± 13 49 ± 9 < 0.001 While cardiovascular outcomes were not evaluated, the present SGOT (≥ 40 IU/L) (%) 34 35 1.0 study suggests that the beneficial effects of alcohol may be relat- SGPT (≥ 40 IU/L) (%) 24 14 0.48 ed to a reduced atherosclerotic plaque burden. Prior studies of alcoholics or heavy drinkers are lacking. In an Sensitivities, specificities, negative, and positive predictive val- autopsy study, Thompson et al. found less CAD in alcoholics as ues of elevated SGPT (≥ 40 IU/L) and SGOT (≥ 40 IU/L) values compared to nonalcoholic men but similar degrees of aortic ath- for the identification of left ventricular dysfunction are shown in erosclerosis.9 Therefore, the coronary protective effect of alcohol Table 2. In both the alcoholic and nonalcoholic groups, the does not seem to diminish even with heavy alcohol consumption. patients who had normal LVEF were more likely to have normal The prevalence of traditional cardiovascular risk factors for SGPT measurements (87% vs. 14%; p < 0.05). For the entire CAD was similar in the two groups with the exception of dia- group, serum SGPT levels correlated inversely with LVEF (r = betes. There were also no significant differences in the mean -0.21; p < 0.05). The prevalence of SGPT levels and LVEF in concentrations of total cholesterol, HDL, LDL, triglycerides the alcoholic and nonalcoholic groups are shown in Figure 2. between the two groups. This would suggest that the protective In a stepwise multivariate logistic analysis, alcoholic status was effects of alcoholism in the prevalence and severity of CAD also found to be an independent predictor of CAD (r = -0.15; p < may be unrelated to lipids. 0.05). Univariate predictors of CAD and left ventricular The alcoholic patients who have developed cirrhosis of the dysfunction were entered into the multivariate model. In a step- liver may have developed decreased testosterone levels through wise multivariate logistic analysis using, alcohol status, diabetes, the destruction of their hepatocytes.10 The decrease in testosterone SGPT, and CAD, the independent predictors of left ventricular levels in the alcoholic group who developed cirrhosis may directly dysfunction were alcohol status and SGPT (r = 0.24; p < 0.05). contribute to the increase in CAD in the alcoholic patients with

Table 2. Sensitivities and specificities in predicting patients with ejection fraction < 40% identified by abnormally high SGOT and SGPT enzyme levels.

Positive Predictive Negative Predicive Enzyme Population Sensitivity (%) Specificity (%) Value (%) Value (%) p-value SGOT Alcoholic 38 68 41 65 0.66 ≥ 40 IU/L Nonalcoholic 42 66 15 89 0.5 All patients SGOT ≥ 40 IU/L 39 67 24 81 0.45 SGPT Alcoholic 27 77 42 64 0.64 ≥ 40 IU/L Nonalcoholic 43 90 42 93 0.001 All patients SGPT ≥ 40 IU/L 33 86 42 82 0.001

306 The Journal of Invasive Cardiology Effects of Alcoholism on CAD and LV Dysfunction in Male Veterans

Patients admitted to Department of Veterans Affairs New York and underwent cardiac catheteriza- Harbor Healthcare with chest pain from 1994 to 2002 tion after abnormal stress testing. n = 4,908 ICD-9 codes of alcoholic pancreatitis and alcoholic liver cirrhosis were used as surrogates for heavy alcohol inges- Alcoholic patients admitted Nonalcoholic patients with chest pain§ admitted with chest pain tion rather than an actual measure of n = 390 n = 4,518 alcohol consumption. The authors felt that that the presence of end- organ damage from was Alcoholic patients admitted with chest Nonalcoholic patients admitted with chest a surrogate measure of the degree of pain and had positive stress test* pain and had positive stress test* n = 142 n = 890 alcoholism since the actual alcohol consumed was unavailable. Since not all alcoholics experience end-organ Alcoholic patients with chest pain and Nonalcoholic patients with chest pain and damage this may represent the worst had positive stress test undergoing had positive stress test undergoing cardiac case scenario and a skewed patient cardiac catheterization catheterization n = 100 n = 200 population. These results are not nec- essarily applicable to women because

Alcoholic patients with alcoholic liver Alcoholic patients with alcoholic all of the study subjects were male. cirrhosis, chest pain and had positive pancreatitis, chest pain and positive These groups of patients constitute a stress test undergoing cardiac stress test undergoing cardiac small percentage of patients who con- catheterization catheterization n = 91 n = 9 sume alcohol and are subject to the limitations of small groups. Despite § The alcoholic patients had a diagnosis of either alcoholic liver cirrhosis or alcoholic pancreatitis. the limitations, it is concluded that * The stress test was considered positive if the results of the test on the chart were documented as “positive”, “equivocal”, or “nondiagnostic’. there appears to be an inverse relation- ship between CAD and left ventricu- Figure 3. Methodology of the study. lar dysfunction in alcoholic patients. preserved left ventricular function. Testosterone replacement has References been associated with a delay time to , lowering choles- 1. Gaziano M, Buring JE, Breslow JL, et al. Moderate alcohol intake, increased levels of terol, and a lowering serum tumor necrosis factor-alpha, a proin- high-density lipoprotein and its subfractions, and decreased risk of myocardial infarc- 11 tion. N Engl J Med 1993;329:1829–1834. flammatory cytokine. Serum tumor necrosis factor alpha has 2. Fruchart JC, Ailhaud G, Bard JM. Heterogeneity of high density lipoprotein parti- been found to be elevated in patients with liver cirrhosis.12 cles. Circulation 1993;87(4 Suppl):III22–III127. Oxidative stress and inflammation such as elevated levels of 3. Gill JS, Shipley MJ, Tsementzis SA, et al. Alcohol consumption — A risk factor for hemorrhagic and non-hemorrhagic stroke. Am J Med 1991;90:489–497. tumor necrosis factor-alpha are associated with the propagation 4. Piano MR. Alcoholic cardiomyopathy: Incidence, clinical characteristics, and of coronary artery disease and congestive .13–15 As com- pathophysiology. Chest 2002;121:1638–1650. 5. McKelvie R. Heart failure. Clin Evid 2004;6:122–148. pared to nonalcoholic patients, alcoholic patients showed a higher 6. Deutscher S, Rockette HE, Krishnaswami V. Evaluation of habitual excessive alco- prevalence of left ventricular dysfunction and a lower mean LVEF. hol consumption on myocardial infarction risk in coronary disease patients. Am Heart J 1984;108(4 Pt.1):988–995. The findings are consistent with other studies showing an associa- 7. Mukamal KJ, Ascherio A, Mittleman MA, et al. Alcohol and risk for ischemic tion between heavy alcohol consumption and dilated car- stroke in men: The role of drinking patterns and usual beverage. Ann Intern Med diomyopathy.4 Although diabetes was more common among 2005;142:11–19. 8. Berger K, Ajani UA, Kase CS, et al. Light-to-moderate alcohol consumption and alcoholics and has been associated with dilated cardiomyopa- the risk of stroke among U.S. male physicians. N Engl J Med 1999;341:1557–1564. thy, diabetes was not an independent predictor of left ventricu- 9. Thomsen JL. Atherosclerosis in alcoholics. Ugeskr Laeger 1997;159:757–760. 10. Kovacs EJ, Messingham KA, Influence of alcohol and gender on immune response. lar dysfunction on multivariate analysis in the present study. Alcohol Res Health 2002;26:257–263. A novel finding is the high negative predictive value of a nor- 11. Malkin CJ, Pugh PJ, Morris PD, et al. Testosterone replacement in hypogonadal men with improves ischaemic threshold and quality of life. Heart mal SGPT in excluding left ventricular dysfunction. It has been 2004;90:871–876. suggested that patients develop either alcoholic related cardiomy- 12. Lin SY, Wang YY, Sheu WH. Increased serum leptin concentrations correlate with opathy or cirrhosis in a mutually exclusive manner.16 To date, no soluble tumour necrosis factor receptor levels in patients with cirrhosis. Clin Endocrinol (Oxf) 2002;57:805–811. study has evaluated specific serum liver enzymes as a marker of 13. Zhang C, Hein TW, Wang W, et al. Activation of JNK and xanthine oxidase by left ventricular dysfunction. Although elevated serum SGPT and TNF-alpha impairs nitric oxide-mediated dilation of coronary arterioles. J Mol Cell Cardiol 2006;40:247–257. SGOT levels do not appear to be useful in differentiating patients 14. Lucas DL, Brown RA, Wassef M, Giles TD. Alcohol and the cardiovascular system. with or without left ventricular dysfunction, normal SGPT values Research challenges and opportunities. J Am Coll Cardiol 2005;45:1916–1924. 15. Odeh M, Sabo E, Oliven A. Circulating levels of tumor necrosis factor-alpha corre- have a high negative predictive value and therefore are useful in late positively with severity of peripheral oedema in patients with right heart failure. excluding left ventricular dysfunction. Eur J Heart Fail 2006;8:141–146. Epub 2005 Study limitations. This study is subject to the inherent 16. Ahmed SS, Howard M, ten Hove W, et al. Cardiac function in alcoholics with cir- rhosis: Absence of overt cardiomyopathy — Myth or fact? J Am Coll Cardiol limitations of a retrospective analysis. All patients were male 1984;3:696–702.

Vol. 18, No. 7, July 2006 307