ORIGINAL INVESTIGATION. Opposite Associations of Carbohydrate-Deficient Transferrin and -Glutamyltransferase With Prevalent Coronary Heart Disease

Transcription

1 ORIGINAL INVESTIGATION Opposite Associations of Carbohydrate-Deficient Transferrin and -Glutamyltransferase With Prevalent Coronary Heart Disease Pekka Jousilahti, MD, PhD; Erkki Vartiainen, MD, PhD; Hannu Alho, MD, PhD; Kari Poikolainen, MD, PhD; Pekka Sillanaukee, PhD Background: Moderate alcohol consumption may reduce the risk of coronary heart disease (CHD), but excessive alcohol consumption is probably harmful to the heart. We analyzed the association of 2 commonly used markers of alcohol consumption carbohydrate-deficient transferrin (CDT) and -glutamyltransferase (GGT) and self-reported alcohol consumption with prevalent CHD. Methods: The study included a random sample of 3666 Finnish men aged 25 to 74 years who participated in a risk factor survey in The cross-sectional association of CDT, GGT, and self-reported drinking with CHD was analyzed by means of logistic regression models. Results: The CDT level was inversely and GGT level positively associated with CHD risk. The odds ratios (adjusted for age, smoking, total and high-density lipoprotein cholesterol levels, systolic blood pressure, and body mass index) of CHD among men in the fourth quartiles of CDT and GGT, as compared with the first quartiles, were 0.69 and 1.76, respectively. In a composite risk assessment, men with normal CDT levels ( 20 U/L) and elevated GGT levels ( 80 U/L) had nearly 8-fold adjusted risk of CHD as compared with the men with normal GGT levels and elevated CDT levels. Self-reported alcohol consumption had an inverse association with CHD risk, which disappeared after adjustment for the other risk factors. Conclusions: Levels of CDT and GGT may be indicators of factors behind the curvilinear association between alcohol consumption and CHD risk. The CDT level seems to be related to beneficial biological changes and GGT level with the changes that are detrimental to the cardiovascular system. The inverse association of CDT level with CHD risk will be examined further in a forthcoming prospective study. Arch Intern Med. 2002;162: From the Department of Public Health (Dr Jousilahti) and Research Unit of Substance Abuse Medicine (Dr Alho), University of Helsinki, Helsinki; Department of Epidemiology and Health Promotion (Drs Jousilahti and Vartiainen) and Department of Mental Health and Alcohol Research (Dr Alho), National Public Health Institute, Helsinki; Finnish Foundation for Alcohol Studies, Helsinki (Dr Poikolainen); Oy Finnish Immunotechnology Ltd, Tampere (Dr Sillanaukee); and Department of Medical Biochemistry and Research Unit, Tampere University and University Hospital, Tampere (Dr Sillanaukee), Finland. THERE IS reasonably good evidence that light to moderate alcohol consumption may reduce the risk of coronary heart disease (CHD) as compared with total abstinence, but that excessive alcohol consumption is probably detrimental to the heart. 1-4 Even though this J-shaped association between alcohol consumption and CHD risk has been observed in several epidemiologic studies, there still is controversy concerning the limit for safe drinking. 5 The definition of safe limits is complicated by the fact that self-reported alcohol intake is usually underestimated. 5,6 Because of this, biological markers might be useful in the evaluation of the risk of CHD related to alcohol consumption. Carbohydrate-deficient transferrin (CDT) and -glutamyltransferase (GGT) are commonly used markers of alcohol consumption. 7,8 High levels of GGT have been shown to be associated with increased CHD risk. 9 The association between CDT levels and the risk of CHD has not been studied before, to our knowledge. A previous study, however, has shown that CDT is associated with a nonatherogenic serum lipid profile and GGT with an atherogenic profile. 10 The aim of the present study was to analyze the association of CDT and GGT, as well as selfreported alcohol consumption, with the prevalence of CHD. RESULTS All of the analyzed markers (CDT, GGT, and self-reported alcohol consumption) had a positive association with smoking prevalence (Table 1). The CDT level had 817

2 SUBJECTS AND METHODS To assess the levels of behavioral and biological healthrelated factors among the population, a large cross-sectional risk factor survey was performed in Finland in The survey was conducted according to the World Health Organization MONICA (Monitoring Trends and Determinants of Cardiovascular Disease) protocol in 5 geographical areas: the Helsinki-Vantaa region in southern Finland, the Turku-Loimaa region in southwestern Finland, Kuopio and North Karelia provinces in eastern Finland, and the Oulu Province in northern Finland. In each study area, an age- and sex-stratified random sample of 2000 subjects was drawn from the population of people between the ages of 25 and 64 years. In addition, a sample of 250 women and 500 men aged 65 to 74 years was drawn in North Karelia and the Helsinki-Vantaa region. The total sample size was 5500 women and 6000 men. Of them, 4193 women (76%) and 4254 men (71%) participated. Because of the relatively small number of women with CHD, only men were included in the present study. A total of 372 men were excluded because of missing data on CDT, GTT, or serum cholesterol levels; self-reported alcohol consumption; smoking; blood pressure; or body mass index (BMI). Two hundred sixteen men were excluded from the analyses because of the discrepancy between self-reported and registered data on the occurrence of CHD or because of self-reported stroke. Thus, 3666 men were included in the analyses. The study was approved by the ethical committee of the National Public Health Institute, Helsinki, Finland, and the subjects gave informed consent. Data on smoking, alcohol consumption, and the occurrence of symptoms of CHD were obtained from a selfadministered questionnaire. Quantitative estimation of alcohol intake was based on a set of structured questions inquiring about the amount and frequency of drinking during the past year. Average weekly alcohol consumption was calculated on the assumption that a standard unit of beer, mixed drinks, spirits, and wine contained 12 g of alcohol each, whereas a unit of cider or coolers contained 4 g of alcohol. At the study site, specially trained nurses measured blood pressure, weight, and height by using a standardized protocol. Weight was measured in light clothing and height without shoes. The BMI (weight in kilograms divided by the square of height in meters) was used as an indicator for obesity. A venous blood sample was drawn and centrifuged within 1 hour. Serum samples were transported at room temperature to the central laboratory, where GGT, serum total cholesterol, and high-density lipoprotein (HDL) cholesterol were analyzed within a few days. Serum for CDT determination was stored at 70 C a few months before the analyses. Serum CDT levels were analyzed by a double antibody kit (CDTect; Pharmacia & Upjohn Diagnostics AB, Uppsala, Sweden) according to the manufacturer s instructions. Levels of GGT were measured by a kinetic method (Oy Medix Biochemica AB, Kauniainen, Finland) based on the recommendation of the European Committee for Clinical Laboratory Standards. Serum total and HDL cholesterol levels were determined by an enzymatic method (CHOD-PAP; Boehringer Mannheim, Mannheim, Germany). Data on the occurrence of CHD were obtained from the National Social Insurance Institution s register of subjects entitled to special reimbursement for CHD drugs. To receive the special reimbursement, the diagnosis of CHD is usually made by a specialist in internal medicine or a cardiologist, and the diagnosis needs to be based on symptoms and clear ischemic changes in the electrocardiogram, either at rest or during an exercise test, or findings on coronary angiography. The statements documenting these findings are then reviewed and accepted by the expert physicians of the National Social Insurance Institution. Subjects who did not receive the special reimbursement but who reported having CHD or stroke on the questionnaire were excluded from the analysis. Thus, the subjects classified as having CHD probably had the disease and those who were classified as disease free probably did not have clinically significant atherosclerosis. The associations of CDT, GGT, and self-reported alcohol consumption with the prevalence of CHD were analyzed by logistic regression models. In the analyses, the first (ie, lowest) quartiles of CDT and GGT were used as the reference categories, and the other quartiles were included in the model as dummy variables. In self-reported alcohol consumption, nondrinkers were used as a reference category. The odds ratios (ORs) and 95% confidence intervals (CIs) of prevalent CHD are presented for each quartile with a test for linear trend across the quartiles. The composite risk assessment was done by dichotomizing the variables by means of the recommended reference values (20 U/L for CDT and 80 U/L for GGT) and analyzing the combined risk in the resulting 4 categories with the help of dummy variables. All analyses were first adjusted for age only and then further for age, smoking, serum total and HDL cholesterol levels, blood pressure, and BMI. The associations of CDT, GGT, and self-reported alcohol consumption with the known cardiovascular risk factors were also examined. a positive association with HDL cholesterol level and an inverse association with BMI, but CDT level was not associated with total cholesterol level or systolic blood pressure. The GGT level had a positive association with total cholesterol level, systolic blood pressure, and BMI, and an inverse association with HDL cholesterol level. Selfreported alcohol consumption had a positive association with serum total and HDL cholesterol levels, systolic blood pressure, and BMI. The correlation coefficients of CDT and GGT levels with self-reported alcohol consumption were 0.32 (P.001) and 0.24 (P.001), respectively. Their correlation with each other was 0.19 (P.001). The CDT level had a progressive inverse association and GGT level had a progressive positive association with CHD risk (Table 2). The ORs of CHD among men in the second, third, and fourth quartiles of CDT, as compared with the first quartile, were 0.83, 0.67, and 0.50 (P value for trend,.001). The respective ORs in the second, third, and fourth quartiles of GGT were 1.05, 1.36, and 1.67 (P value for trend,.005). The inverse association of CDT and CHD decreased somewhat after ad- 818

3 Table 1. Age-Adjusted Risk Factor Levels by Quartiles of CDT and GGT, and in Categories of Self-reported Alcohol Consumption* Quartile/Category First Second Third Fourth P Value CDT, U/L Smoking, % Total cholesterol, mg/dl HDL cholesterol, mg/dl Systolic blood pressure, mm Hg Body mass index, kg/m GGT, U/L Smoking, % Total cholesterol, mg/dl HDL cholesterol, mg/dl Systolic blood pressure, mm Hg Body mass index, kg/m Self-reported weekly alcohol consumption, g Smoking, % Total cholesterol, mg/dl HDL cholesterol, mg/dl Systolic blood pressure, mm Hg Body mass index, kg/m *CDT indicates carbohydrate-deficient transferrin; GGT, -glutamyltransferase; and HDL, high-density lipoprotein. To convert cholesterol to millimoles per liter, multiply by Table 2. Odds Ratios (ORs) and 95% Confidence Intervals (CIs) of Coronary Heart Disease by Quartiles of Carbohydrate-Deficient Transferrin (CDT) and -Glutamyltransferase (GGT), and in Categories of Self-reported Alcohol Consumption Quartile/Category Second Third Fourth First, OR OR 95% CI OR 95% CI OR 95% CI P Value for Trend CDT Model 1* Model GGT Model Model Self-reported alcohol consumption Model Model *Adjusted for age. Adjusted for age, smoking, serum total and high-density lipoprotein cholesterol levels, systolic blood pressure, and body mass index. In self-reported alcohol consumption, the reference group includes only total abstainers. justment for smoking, total and HDL cholesterol levels, systolic blood pressure, and BMI. The positive association between GGT and CHD changed only slightly after adjustment for the other risk factors. Self-reported alcohol consumption had a borderline inverse association with CHD risk, which, however, disappeared after adjustment for the other risk factors. The OR of CHD among men with elevated CDT level ( 20 U/L) and normal GGT level ( 80 U/L) was 0.22 compared with men who had normal levels for both values (Table 3). The OR among men with elevated GGT level and normal CDT level was When both CDT and GGT levels were elevated, the risk did not differ from the reference category. When the men with elevated GGT level and normal CDT level were compared with men with elevated CDT level and normal GGT level, the ageadjusted OR was (95% CI, ). Adjustment for the other risk factors decreased the OR to 7.74 (95% CI, ). COMMENT In the present study, serum CDT and GGT levels, which were regarded as biological markers of alcohol consumption, had opposite associations with the prevalence of CHD. The risk of CHD decreased with increasing CDT level but increased with increasing GGT level. The CDT and GGT levels also had contrasting associations with major cardiovascular risk factors. High GGT level was associated with high levels of all risk factors. Either the association of CDT with the known cardiovascular risk factors was neutral, or high CDT levels were related to low risk factor levels. In the composite risk assessment, men with elevated GGT level and normal CDT level had nearly 8-fold risk (adjusted for 819

4 Table 3. Odds Ratios (ORs) and 95% Confidence Intervals (CIs) of Coronary Heart Disease by Combinations of Carbohydrate-Deficient Transferrin (CDT) and -Glutamyltransferase (GGT) Levels Combination* B C D A, OR OR 95% CI OR 95% CI OR 95% CI Model Model *A indicates CDT of 20 U/L or less and GGT of 80 U/L or less (reference category); B, CDT greater than 20 U/L and GGT of 80 U/L or less; C, CDT of 20 U/L or less and GGT greater than 80 U/L; and D, CDT greater than 20 U/L and GGT greater than 80 U/L. Adjusted for age. Adjusted for age, smoking, serum total and high-density lipoprotein cholesterol levels, systolic blood pressure, and body mass index. the other risk factors) as compared with the men with elevated CDT level and normal GGT level. Self-reported alcohol consumption had an inverse association with CHD, which, however, disappeared after adjustment for the other risk factors. The finding is at least partly explained by the cross-sectional study design and changes in drinking behavior due to an existing disease. According to the self-reporting, alcohol consumption among the study population was also relatively low. This is probably partly due to underreporting among the responders, and partly because the heaviest drinkers do not usually participate in surveys. Several studies have shown that moderate alcohol consumption decreases the risk of CHD, as compared with total abstinence. 1-4 The primary mechanisms proposed for the cardioprotective effect of alcohol are that alcohol has positive effects on lipid metabolism and on the hemostatic system. It has been shown that moderate drinkers have higher HDL cholesterol level than nondrinkers, and the HDL cholesterol increasing effect of alcohol has also been observed in clinical trials. 11,12 A polymorphism in the gene for alcohol dehydrogenase type 3 alters the rate of alcohol metabolism. A recent nested casecontrol study demonstrated that moderate drinkers who were homozygous for the slow-oxidizing alcohol dehydrogenase type 3 allele had higher HDL cholesterol levels and a substantially decreased risk of CHD. 13 It has also been demonstrated that alcohol has antithrombotic effects. 11,14 On the other hand, even relatively small amounts of alcohol have been shown to increase blood pressure. 15 It has been debated whether the beneficial effect of moderate alcohol consumption on the heart is due to ethanol or some other substances in alcoholic beverages. 2 Some researches have suggested that red wine is particularly healthful, but the issue is still controversial. The protective effect of alcohol may also be associated with drinking pattern. 16 A positive association between GGT and the risk of CHD has been reported previously. 9 The present study is the first one, to our knowledge, in which the association of CDT with CHD risk has been analyzed. We can only speculate on the mechanisms for the opposite associations of CDT and GGT with CHD. Only part of it can be explained by the known cardiovascular risk factors, particularly lipids. It is possible that CDT is a marker for earlier phases of alcohol consumption, and GGT might reflect toxic effects of ethanol on hepatic lipid metabolism. It has also been shown that CDT is a more specific marker of alcohol consumption than GGT. 17 It has been suggested that CDT level reflects drinking frequency, whereas GGT level is more influenced by drinking intensity. 18 The correlation of CDT and GGT levels with selfreported alcohol consumption was higher than their correlation with each other, suggesting that the responses of CDT and GGT levels to alcohol consumption occur by different mechanisms. 19 According to the self-reported alcohol consumption records, most of the subjects were drinking at relatively low levels. This is the range where the J-shaped curve relating alcohol intake to CHD first declines and then begins to rise again. The usual explanation for such a biphasic curve is that there are 2 or more factors involved. Our results suggest the possibility that CDT and GGT are indicators of these factors, with CDT being related somehow to beneficial factors from alcohol consumption and GGT to the alcohol factors that are detrimental to the cardiovascular system. Other factors in addition to alcohol consumption may be involved in the association of CDT and GGT with CHD risk. For example, obesity increases GGT levels and is also a risk factor for CHD. 20 In our study, however, the positive association between GGT and CHD remained after adjustment for the other risk factors, including BMI. Some studies have suggested that high hemoglobin level may be a risk factor for CHD. 21 The CDT level is influenced by iron metabolism, and it increases in iron deficiency. 22,23 Thus, variation in iron intake may be a confounding factor in the association between CDT and CHD risk. Because of the cross-sectional study design, we cannot yet draw any causal conclusions on the association between CDT and CHD risk. The other limitation of our study is that it included only men. Most participants were light to moderate drinkers. Therefore, our findings may be limited to male social drinkers. We conclude that CDT and GGT levels have opposite associations with CHD risk. It is possible that CDT and GGT levels tend to separate the factors that are responsible for the positive and negative cardiovascular effects of alcohol consumption. This novel observation may lead to better understanding of the complex relation of alcohol to CHD risk. However, the inverse association of CDT level with CHD risk is a preliminary finding and will be studied further in a forthcoming prospective study among the CHD-free subjects in the present cohort. 820

5 Accepted for publication July 31, Corresponding author and reprints: Pekka Jousilahti, MD, PhD, University of Helsinki, Department of Public Health, PO Box 41, FIN University of Helsinki, Finland ( REFERENCES 1. Thun MJ, Peto R, Lopez AD, et al. Alcohol consumption and mortality among middle-aged and elderly U.S. adults. N Engl J Med. 1997;337: Rimm EB, Klatsky A, Grobbee D, Stampfer MJ. Review of moderate alcohol consumption and reduced risk of coronary heart disease: is the effect due to beer, wine, or spirits? BMJ. 1996;312: Doll R. One for the heart. BMJ. 1997;20: Poikolainen K. It can be bad for the heart, too: drinking patterns and coronary heart disease. Addiction. 1998;93: Kalant H, Poikolainen K. Moderate drinking: concepts, definitions and public health significance. In: MacDonald I, ed. Health Issues Related to Alchohol Consumption. 2nd ed. Oxford, England: ILSI Europe & Blackwell Science; 1999: Midanik L. The validity of self-reported alcohol consumption and alcohol problems: a literature review. Br J Addict. 1982;77: Stibler H. Carbohydrate-deficient transferrin in serum: a new marker of potentially harmful alcohol consumption reviewed. Clin Chem. 1991;37: Sillanaukee P. Laboratory markers of alcohol abuse. Alcohol Alcohol. 1996;31: Wannamethee G, Ebrahim S, Shaper AG. Gamma-glutamyltransferase: determinants and association with mortality from ischemic heart disease and all causes. Am J Epidemiol. 1995;142: Nikkari ST, Koivu TA, Anttila P, Raunio I, Sillanaukee P. Carbohydrate-deficient transferrin (CDT) and gamma-glutamyltransferase (GGT) are inversely associated with lipid markers of cardiovascular risk. Eur J Clin Invest. 1998;28: Rimm EB, Williams P, Fosher K, Criqui M, Stampfer MJ. Moderate alcohol consumption and lower risk of coronary heart disease: meta-analysis of effects on lipids and haemostatic factors. BMJ. 1999;319: Paunio M, Heinonen OP, Virtamo J, et al. HDL cholesterol and mortality in Finnish men with special reference to alcohol intake. Circulation. 1994;90: Hines LM, Stampfer MJ, Ma J, et al. Genetic variation in alcohol dehydrogenase and the beneficial effect of moderate alcohol consumption on myocardial infarction. N Engl J Med. 2001;344: Hamsten A. The hemostatic system and coronary heart disease. Thromb Res. 1993;70: Klatsky AL. Alcohol and hypertension. Clin Chim Acta. 1996;246: McElduff P, Dobson AJ. How much alcohol and how often? population based case-control study of alcohol consumption and risk of a major coronary event. BMJ. 1997;314: Burke V, Puddey IB, Rakie V, et al. Carbohydrate-deficient transferrin as a marker of change in alcohol intake in men drinking 20 to 60 g of alcohol per day. Alcohol Clin Exp Res. 1998;22: Anton RF, Stout RL, Roberts JS, Allen JP. The effect of drinking intensity and frequency on serum carbohydrate-deficient transferrin and gamma-glutamyl transferase levels in outpatient alcoholics. Alcohol Clin Exp Res. 1998;22: Randell E, Diamandis EP, Goldberg DM. Changes in serum carbohydratedeficient transferrin and gammaglutamyl transferase after moderate wine consumption in healthy males. J Clin Lab Anal. 1998;12: Robinson D, Whitehead TP. Effect of body mass and other factors on serum liver enzyme levels in men attending for well population screening. Ann Clin Biochem. 1989;26: Salonen JT, Nyyssönen K, Korpela H, Tuomilehto J, Seppänen R, Salonen R. High stored iron levels are associated with excess risk of myocardial infarction in eastern Finnish men. Circulation. 1992;86: De Feo TM, Fargion S, Duca L, et al. Carbohydrate-deficient transferrin, a sensitive marker of chronic alcohol abuse, is highly influenced by body iron. Hepatology. 1999;29: Sorvajärvi K, Blake JE, Israel Y, Niemelä O. Sensitivity and specificity of carbohydrate-deficient transferrin as a marker of alcohol abuse are significantly influenced by alterations in serum transferrin: comparison of two methods. Alcohol Clin Exp Res. 1996;20: