Correlations of serum lipids and lipoproteins with gamma-glutamyltransferase and attitude to alcohol consumption *

Transcription

1 Ann Clin Biochem 1982; 19: Correlations of serum lipids and lipoproteins with gamma-glutamyltransferase and attitude to alcohol consumption * G FEX, H KRSTENSON AND E TRELL From the DepartmentofClinical Chemistry, DepartmentofAlcohol Diseases and the DepartmentofPreventive Medicine, University oflund, Malmo General Hospital, S Malmo, Sweden SUMMARY The correlation of three markers of alcohol consumption: serum concentration of gamma-glutamyltransferase (GT), high density lipoprotein cholesterol (HDL-cholesterol), apolipoprotein A (Apo A), and the results of a questionnaire (Mm-MAST) designed to measure the attitude towards alcohol consumption were determined in a population of healthy men aged 48 years. Apo A (r = 0,27) and GT (r = 0,1) were correlated (p < 0 001) to the questionnaire score to a similar degree while the correlation between HDL-cholesterol and the questionnaire score was (r = 0,18), somewhat less good but statistically significant (p < 0 01). There was no correlation between GT and HDL-cholesterol or Apo A. Heavy alcohol consumption in chronic alcoholics is associated with increased levels of high density lipoproteins (HDL) in plasma. 1 - n these chronic alcoholics plasma HDL-cholesterol and apolipoprotein A (Apo A) concentrations were elevated as frequently as gamma-glutamyltransferase (GT) and it was suggested that HDL-cholesterol or Apo A might be used as indicators of heavy alcohol consumption in the same way as GT.4-8 Questionnaires have also been used in screening for alcoholism in various populations.prp' To test the possible value of HDL-cholesterol and Apo A as markers of alcohol consumption in a population of healthy men in contrast to one group of chronic alcoholics- we have determined HDLcholesterol, Apo A, GT as well as total cholesterol and triglycerides and correlated these variables to one another and to the number of yes-answers obtained on a questionnaire (Mm-MAST) designed to measure the attitude to alcohol.p Materials and methods Three hundred men aged 48 years were investigated. All were subjectively healthy and had accepted an invitation to a general medical screening investigation free of charge. The details of this health survey have been given." n conjunction with the screening * Supported by grant 0X-064 from the Swedish Medical Research Council, and from the Swedish Delegation for Social Research, Ministry of Health and Social Affairs. Table 1 The Malmo modification of the Michigan alcoholism screening test (Mm-MAST) t Do you take a drink before going to a party? 2 Do you usually drink a bottle of wine or corresponding amounts of alcohol over the week-end? Do you drink a couple of drinks (or beers) a day to relax? 4 Do you tolerate more alcohol now than you did ten years ago? 5 Have you difficulties not drinking more than your friends? 6 Do you fall asleep after moderate drinking without knowing how you got to bed? 7 Do you have a bad conscience after drinking? B Do you take a drink (a beer) the day after a party? 9 Do you try to avoid alcoholic beverage for a determined period of time. eg, a week? investigation all individuals answered an extensive questionnaire of 260 questions, including nine questions about drinking habits (Mm-MAST).1 These questions are listed in Table 1. Fastingblood samples were obtained in the morning. Serum was obtained by centrifugation. GT,14 triglyceride.p total cholesterol,16 and HDL-cholestero1 were determined within 48 hoursand the samples were kept at +4 C until analysed. Apo A was determined by electroimmunoassay on samples which had been kept at -20 C for a maximum of two weeks as described previously" with pooled serum as standard. The results were expressed as the percentage of the value of the pool." The answer to the nine questions about drinking habits were evaluated as the number of yes-answers to the questions. The higher the figure, the more permissive the attitude to alcohol.p 45

2 46 Fex, Kristenson, and Trell Table 2 Correlation coefficients between plasma cholesterol (Chot), triglyceride (TG), H'Dli-cholesterol (HDL Chol), Apolipoprotein A (Apo A), glutamyltransferase (GT), and score Chol TG HDL-Chol ApoA GT Score p < 0 01 p < Chal TG H DL-Chal O S Apa A GT Score Results The linear correlation coefficientsare shown in Table 2 and the quintile distributions of serum GT, HDLcholesterol, and Apo A in Figures 1 to. As there were few individuals with high questionnaire scores this variable was divided into classes according to score instead of quintiles (Fig. 4). n each quintile (class) the number of individuals with 'abnormal' values are indicated. 'Abnormal' limits were > 1 {J.kat{l for GT (the upper limit for LOO 4 <r: o u...;z: ioo ::;: (5 E..s e"' Ouintiles of GT 1 Fig. Serum concentration ofhdl-cholesterol and Apo A and questionnaire score in quintiles of the GT distribution. The number of'obnormal' values, ie, HDls-cholesterol and Apo A ;;;. 95th percentile and score» 2, are given. Mean and standard deviations for GT are given. iii (5 "' J:: u...j Quintiles of GT Quintiles of HDL- cholesterol Fig. 2 Serum concentrations of Apo A and GTand questionnaire score in quinttles of the HDL-cholesterol distribution. The number ofabnormal values, ie, Apo A ;;;. 95th percentile and GT;;;. ] 0 ukat]! and score ie 2, are given. Mean and standard deviation for HDL-cholesterol. o Quintiles of HOL- cholesterol

3 Correlations of serum lipids and lipoproteins with gamma-glutamyltronsferase ""5 E..se 1 5 iii 15 s: u ' o Fig. Serum concentration ofhdl-cholesterol and GT and questionnaire score in quintiles of the Apo A distribution. The number of abnormal values, ie, 95th percentile for HDL-cholesterol and vkat]l for GT and 2 for score, are given. Mean and standard deviation for Apo A Quintiles of Apo Al 2 5 ' E '0 ' 5..s.J< - l:l. *100 0 s: ':ii' &. 100 « Classes of score u Classes of score Fig. 4 Serum concentration ofhdl-cholesterol, Apo A, and GT in classes ofthe questionnaire score. The number of abnormal values, ie, 95th percentile for HDL-cholesterol and Apo Aand 1 0 ukatllfor G''; are given. The percentage of individuals in each score class was: < 1: 4%; < 2: 26%; < : 14% ;;. : 17%. Mean and standard deviation for score in each class. GT for this age and sex at the local laboratory), 140% for Apo A, and 1 80 mmol/l for HDLcholesterol (95th percentile limits), and 2 for questionnaire score (the limit used in ref. 1). The questionnaire score was significantly correlated to Apo A, GT, and HDL-cholesterol, more strongly to the two former than to the latter, Apo A and HDL-cholesterol were strongly correlated to each other but neither of them was correlated to GT. The partial correlation coefficient of GT and Apo A corrected for questionnaire score was The multiple correlation coefficient between GT, Apo A, and questionnaire score was Questionnaire score increased with increasing GT (Fig. 1). The number of individuals with questionnaire score 2 increased to 50 % of the individuals in the highest GT quintile. The numberof individuals with values 95th percentile for Apo A and HDLcholesterol showed a somewhat more irregular distribution between quintiles of GT. Abnormal GT values ( 1 0 flkat/l) showed a tendency towards a bimodal distribution in quintiles of HDL-cholesterol (Fig. 2) which, however, was less pronounced in quintiles of Apo A (Fig. ). Questionnaire score increased with increased HDLcholesterol and Apo A. The number of individuals with questionnaire score 2 was relatively similar in quintiles of HDL-cholesterol (Fig. 2) but increased to slightly more than 50% of the individuals in the highest quintile of Apo A (Fig. ).

4 48 GT, Apo A, and HDL-cholesterol all increased with increasing questionnaire score (Fig. 4). The relative number of individuals with GT 1 0!J.kat/l was highest in the questionnaire score classes 2. The same tendency was evident for Apo A and, to a lesser extent, for HDL-cholesterol. Discussion Questionnaires have been widely used in screening for alcoholism.v-" Depending on the design of the questionnaire, the way it is administered and the definitions of alcoholism, there is much variation in the reported sensitivity and specificity Questionnaires of the design used in the present investigation (Table 1) seem to have reasonable sensitivity and specificity. S t might, therefore, be assumed that the results obtained with a questionnaire of this type would correlate well with other markers of heavy alcohol consumption like GT,4-8 HDL-eholesterol, and Apo AJ.l-s Surprisingly, GT and Apo A showed about equally strong correlation with the questionnaire score while the correlation between the latter and HDL-cholesterol was less good. The serum Apo A concentration is increased in chronic alcoholics after a period of abuse 1-S and decreases during abstinence. ts potential as a marker for heavy drinking in apparently healthy men has not been explored. This may be worthwhile as Apo A is not correlated with GT (partial correlation coefficient after correction for score was 0.(08) and therefore might be elevated in heavily drinking individuals who have normal GTvalues. The increase of serum GT in heavy alcohol consumptionvf has been claimed to be a result of microsomal enzyme induction by ethanol in the liver.p The mechanisms responsible for the release of GT from the liver into plasma are, however, not understood. Also, not all heavily drinking individuals have high serum GT. The increased HDL concentration may be related to alcohol in another way. t has also been suggested that increased hepatic triglyceride secretion is related to microsomal enzyme induction.!" t is presently believed that a large part of the HDL is generated in plasma from the surface coat of the triglyceride-rich lipoproteins during their degradation, a process catalysed by lipoprotein lipase.p' n chronic alcoholics, the activity of this enzyme is increased for unknown reasons.p and this increased lipoprotein lipase activity may thus be the reason for the high HDL concentration in heavy drinkers. A number of individuals with high GT values had low HDL-cholesterol (Fig. 2). The reason for this is unknown. Low HDL-cholesterol has been reported in alcoholics with liver diseases? who may also have Fex, Kristenson, and Trell high GT. High GT is associated with drug metabolism-" which may not influence HDL-cholesterol concentration. As increases in serum GT and HDL concentration (HDL-cholesterol and/or Apo A) do not seem to occur in the same persons, the above lends further support to the notion that the rises in serum GT and serum HDL concentrations represent different metabolic effects of alcohol. Metabolic consequences of alcohol consumption must be related to the dose, the drinking pattern, and individual factors. However, this makes the picture more complicated and indicates the virtual impossibility of finding one single biochemical marker which detects (with high sensitivity and specificity) individuals who drink heavily. For this purpose questionnaires like the Mm-MAST may be better suited although they are, unfortunately, less easy to use as screening instruments. References 1 Johansson BG, Laurell CoB. Disorders of serum a lipoproteins after alcohol intoxication. Scand J Clin Lab nvest 1969; 2: Johansson BG, Medhus A. ncrease in plasma a lipoproteins in chronic alcoholics after acute abuse. Acta Med Scand 1974; 195: Danielsson B, Ekman R, Fex G, et al. Changes in plasma high density lipoproteins in chronic male alcoholics during and after abuse. Scand J Clin Lab nvest 1978; 8: Rosalki SB. Serum gammaglutamyltranspeptidase. Adv Clin Chern 1975; 17: Rollason JG, Pincherle G, Robinson D. Serum gamrnaglutamyltranspeptidase in relation to alcohol consumption. Clin Chern Acta 1972; 9: Whitehead TP, Clarke CA, Whitfield AGW. Biochemical and haematological markers of alcohol intake. Lancet 1978; 1: Whitfield JB, Hensley WJ, Bryden D, et al. Some laboratory correlates of drinking habits. Ann Clin Biochem 1978; 15: Kristenson H, Trell E, Fex G, et al. Serum ")'-glutamyltransferase: Statistical distribution in a middle-aged male population and evaluation of alcohol habits in individuals with elevated levels. Prev Med 1980; 9: Selzer ML. The Michigan alcoholism screening test. The quest for a new diagnostic instrument. Amer J Psychiatr 1971; 127: Hunt RD, Morse RM, Swensson WM. Diagnosis of alcoholism with a self-administered alcoholism screening test. Mayo Clinic Proc 1980; 55: Sanders WM, Kershaw PW. Screening tests for alcoholism-findings from a community study. Brit J Addict 1980; 75: Screening tests for alcoholism? The Lancet 1980: ii:

5 Correlations ofserum lipids and lipoproteins with gamma-glutamyltransferase 49 1 Kristensson H, Trell E. ndicators of alcohol consumption. Comparisons between a questionnaire (Mm-MAST), interviews and serum y-glutamyltransferase (GGT) in a health survey of middle-aged males. Brit J Addict in press. 14 The Committee on Enzymes of the Scandinavian Society for Clinical Chemistry and Clinical Physiology. Recommended method for the determination of y glutamyltransferase in blood Scand J Clin Lab nvest 1976; 6: Wahlefeld A. Triglyceride determination after enzymatic hydrolysis. n: Bergmeyer HV (ed) Methods of Enzymatic Analysis, 2nd Engl ed, Verlag Chemie, New York: Weinsheim and Academic Press, 1974; ,. Roeschlau P, Berut E, Gruber W. Enzymatische Bestimmungen des Gesarnt-cholesterins im serum. J Klin Chem Klin Biochem 1974; 12: Goldberg DM. The expanding role ofmicrosomal enzyme induction and its implication for clinical chemistry. Clin Chem 1980; 26: ,. Nilsson-Ehle P, Garfinkel AS, Schotz MC. Lipolytic enzymes and plasma lipoprotein metabolism. Ann Rev Biochem : ,. Ekman R, Fex G, Johansson BG, et al. Changes in plasma high density lipoproteins and lipolytic enzymes after long-term heavy ethanol consumption. Scand J Clin Lab nvest. 1981: 41: Devenyi P, Robinson GM, Kapur BM, et al. High density lipoprotein cholesterol in male alcoholics with and without liver disease. AmJMed1981: 71: Acceptedfor publication 14 January 1982