- Email: [email protected]
Menopausal estrogen use, high density lipoprotein cholesterol subfractions and liver function
Atherosclerosis, 49 ( 1983) 3 1-39 Elsevier Scientific Publishers Ireland,
31 Ltd.
ATH 03391
Menopausal Estrogen Use, High Density Lipoprotein Cholesterol Subfractions and Liver Function Jane A. Cauley
‘, Ronald
E. LaPorte ‘, Lewis H. Kuller ‘, Margaret and Rivka Black Sandler *
Bates ’
’Department of Epidemiology, Graduate School of Public Health and ’School of Health Related Professions,
University of Pittsburgh,
Pittsburgh, PA 15261 (U.S.A.)
(Received 13 August, 1982) (Revised, received 27 April, 1983) (Accepted 28 April, 1983)
Summary
Forty-eight menopausal women taking exogenous estrogen were compared with 246 postmenopausal women not on estrogen. The estrogen users had significantly higher total high density lipoprotein (HDL) (76.0 vs 61.4 mg/dl) and HDL, (36.7 vs 23.0 mg/dl) cholesterol than the controls. There was a similar concentration of HDL, cholesterol for the two groups (39.2 for the estrogen users and 38.4 for the controls). A dose-response was evident between the amount of daily estrogen and HDL-total and HDL, cholesterol. The significant differences between the two groups remained after adjusting for body composition, alcohol intake, cigarette smoking and physical activity. There was a significant difference between the two groups in liver function as measured by the liver enzymes, SGOT, SGPT, with liver enzyme concentrations lower in the estrogen users. The results indicate that the increase in the total HDL cholesterol as a result of menopausal estrogen is primarily the result of increased HDL,. The increase could not be explained by alterations in hepatic microsomal activity as measured by liver enzymes since estrogen users had lower concentrations of liver enzymes than non-estrogen users. Key words:
Estrogen treatment - High density zymes - Liver function - Menopause
lipoprotein
cholesterol - Liver
en-
This research was supported in part by a grant from the Alcohol Beverage Medical Research Foundation and Grants AM-l 180 and AM-21 190 from the National Institutes of Health. Reprint requests to Dr. Cauley, Department of Epidemiology, University of Pittsburgh, Pittsburgh, PA 15261, U.S.A.
0021-9150/83/$03.00
0 1983 Elsevier Scientific
Publishers
Ireland,
Ltd.
32
Introduction The relationship of postmenopausal estrogen use to cardiovascular disease is controversial. One report has suggested that postmenopausal estrogen use may be protective against cardiovascular disease [l] whereas others have demonstrated no relationship [2]. Recently, preliminary results from the Lipid Research Clinics (LRC) have suggested an inverse association between estrogen use and total mortality [3]. Yet, the mechanism by which exogenous estrogen can affect risk is not known, but may be through alterations in lipids, in particular the level of high density lipoprotein cholesterol (HDL-C). High density lipoprotein cholesterol is a primary predictor of coronary heart disease (CHD) risk among postmenopausal women [4]. Recently, the heterogeneous nature of the HDL has been recognized, and it has been suggested that the subfraction HDL, cholesterol (HDL,-C) may be a more effective epidemiologic discriminant of CHD risk than the total HDL-C [5]. The research examining the effect of estrogen use on the total HDL-C has consistently reported that estrogen users had higher levels of total HDL-C [6,7]. The effect of postmenopausal estrogen use on the HDL-C subfractions is less clear. Little is known concerning the relationship of estrogen use to HDL-C subfractions. Krauss et al. [8] noted significantly higher levels of HDL, in 11 women, aged 44-66, taking natural estrogens. In support of this, Tikkanen et al. [9] described a 20% increase in the HDL, cholesterol during oral estradiol treatment in 16 women. The failure to control for other known HDL-C determinants such as the degree of obesity, alcohol consumption, physical activity and cigarette smoking, may have influenced their results since there has been some suggestion that estrogen users may differ from non-users on other characteristics other than their lipid values. In the LRC, women on exogenous hormones were found to be thinner [lo]. In addition to failing to control for confounding variables, the sample sizes in both reports were quite small. The current research was designed to investigate the relationship of postmenopausal estrogen utilization to lipoprotein cholesterol concentrations including HDL, and HDL,. In addition, the research controlled for other known determinants of HDL-C in order to evaluate the independent relationship of menopausal estrogen use to the HDL-C subfractions. The mechanisms by which estrogen use is related to increased HDL-C levels have not been fully elucidated. Exogenous estrogen may increase HDL-C levels through an effect on hepatic lipase [ 111. Estrogen may also increase HDL-C through enzyme induction. We, as well as others, have suggested that several determinants of HDL-C, including alcohol consumption, drugs such as phenobarbital, glutethimide, clofibrate, and perhaps physical activity may raise HDL-C levels through increased hepatic microsomal activity [ 12-161. Pynadath and Chanapai have suggested that estrogen has a similar chemical structure as known enzyme inducers [ 171, and thus conceivably could raise HDL-C by the same mechanism as alcohol. We, therefore, examined liver enzymes as indirect measures of liver function in order to determine if the relationship of exogenous estrogen use to HDL-C was possibly the result of increased hepatic microsomal activity or alteration of liver function.
33
Methods Two hundred and ninety-four postmenopausal women, 246 non-estrogen users and 48 estrogen users were recruited as an ancillary study to research investigating the epidemiology of bone loss in postmenopausal women. As part of the research, blood samples were taken for lipoprotein cholesterol analysis and liver enzymes. Upon entry into the study, all women visited the clinic and completed a medical history questionnaire, including current and past hormone use, the Paffenbarger questionnaire for assessment of physical activity which converts activity above and [ 181 and a cigarette and alcohol beyond activities of daily living to kcal/wk questionnaire. Anthropometric measurements were taken, including height, weight, grip strength and the triceps skinfold. The lipid analyses were performed by the Heinz Nutrition Laboratory at the Graduate School of Public Health, University of Pittsburgh. The laboratory has received certification of quality control from the Center for Disease Control. High density lipoprotein cholesterol concentrations were determined by precipitation procedures using heparin and manganese chloride. The subfraction HDL,-C was then precipitated by dextran sulfate. The soluble cholesterol represented HDL,-C [ 191. Cholesterol analyses were performed using the enzymatic method [20]. Results Table 1 presents the baseline characteristics comparing the estrogen users with the non-estrogen users. Both groups were of similar age and menopausal status. The estrogen users had a lower body mass index (BMI) [wt (kg)/ht*(m)], and had higher grip strength. In the estrogen group, all women reported the use of conjugated equine estrogens, but were on different dosages and regimes. Because of the various drug regimes, the dose was summed over a l-month period, and then divided by 30 to estimate a daily estrogen dose. The frequency distribution of daily estrogen dose is presented in Table 2. The majority of the women were on relatively low doses of estrogen. Alcohol consumption was estimated from questionnaire data which asked the frequency and type of beverage consumed. This was converted to ounces of ethanol. Information on cigarette smoking was also obtained from the questionnaire. There were 12 or 23.1% cigarette smokers in the estrogen group and 30 or 12.4% in the non-estrogen group. The lipid profiles for the two groups are presented in Table 3. There was little difference in the total cholesterol and the triglyceride levels. The total HDL and the HDL,-C concentrations were both significantly higher in the estrogen users. The largest difference was in the HDL, subfraction where HDL,-C was 13.8 mg/dl (38%) higher in the estrogen users. There was little difference between the two groups for the HDL, subfraction. The data were also examined for an estrogen dose effect on the HDL-C subfractions. The daily estrogen dose was positively related to the total HDL-C (r = 0.35, P -c O.OOl), and to the HDL,-C (r = 0.41, P < 0.001) in the univariate
34 TABLE
1
BASELINE
CHARACTERISTICS
OF 151 POSTMENOPAUSAL Estrogen users (n = 48)
Non-estrogen (n = 246)
x
(SD_?) 56.6 8.8 47.5
Age (yrs) Years menopausal Age at menopause
1299.5 25.2
(4.0) (23.7)
145.2 63.1 25.2 27.1
(2.8) (3.6) (4.8)
Alcohol consumption (ounces ethanol/day)
0.44
(0.49)
0.36
Daily estrogen
0.37
(0.19)
0.00
dose (mg)
* P < 0.05. X = mean; SD,? = standard
users
(SD.?) 51.6 8.5 49.3
(1027.0)
139.4 64.3 23.1 26.3
Weight (lbs) Height (in) BMI (kg/m*) * Triceps skinfold (mm)
x
(5.3) (5.2) (5.5)
1441.1 21.6
kcal/week (Paffenbarger) Grip strength (kg) *
WOMEN
(4.2) (5.8) (4.6) (1006.7) (4.8) (25.0) (2.2) (3.9) (5.5)
(0.59)
deviation.
analysis. There was no relationship between estrogen dose and HDL,-C, total cholesterol or triglycerides. Since body weight may influence the pharmacologic effect of the drug, the relationship between dose per pound of body weight and HDL-C subfractions was also examined. In this analysis, the estrogen dose per pound continued to be positively related to the total HDL-C (r = 0.28, P < 0.03)
TABLE
2
FREQUENCY MEN Daily conjugated estrogen (mg)
DISTRIBUTION
equine
OF DAILY
ESTROGEN
n
< 0.10 0.1 l-0.20 0.21-0.30 0.3 l-0.40 0.41-0.50 0.5 l-0.60 0.6 l-0.70 0.7 l-0.80 0.8 l-0.90
2 6 13 5 14 3 3 0 0
0.91-1.00
2 4%
DOSE
IN 48 POSTMENOPAUSAL
(%)
(4.2) (12.5) (27.1) (10.4) (29.2) (6.25) (6.25) _ _ (4.2) (loorc]
WO-
35
TABLE LIPID
3 PROFILES
(mg/dl)
-
ESTROGEN
AND NON-ESTROGEN
Triglycerides Total cholesterol HDL-cholesterol HDL,-cholesterol
** **
HDL,-cholesterol
**p < 0.01. X = mean; SD.%?= standard
WOMEN
Non-estrogen (n = 246)
Estrogen (n = 48)
Lipid
POSTMENOPAUSAL
7
(SD?)
x
(SDY)
110.4 229.4 76.0 36.7
(47.0) (35.4) (17.1) (15.6)
113.4 228.0 61.4 23.0
(59.5) (41.2) (14.5) (11.0)
39.2
(7.0)
38.4
(7.4)
deviation.
and to HDL,-C (r = 0.28, P < 0.03), although the magnitude of the correlation coefficient decreased. These findings are graphically presented in Fig. 1 where HDL-total cholesterol and HDL,- and HDL,-C are examined by estrogen dose. Tertile I corresponds to a daily estrogen dosage of 0.04-0.25 mg; Tertile II; 0.26-0.46 mg; Tertile III: 0.47-0.96 mg. A linear relationship is present for both the total HDL-C and the HDL,-C. Since the estrogen users were thinner and there was a tendency for the estrogen users to consume more alcohol and be more active, these variables may have contributed to the increased HDL-total cholesterol and the HDL,-C levels in the estrogen users. Therefore, analysis of co-variance was employed to adjust for potential confounders. Table 4 presents the mean HDL-C levels adjusted for alcohol mo-
/------A
la-
3
a
-5 d 3 J
so-
40-
I
-_______-------& /xX-
/
//
lsged
A BDL-1 C+tOL
0
x HDL-2 CHOL ---0 HDL-3 -__-A- c+iOL
NOBl’ROGD(
lmnl.EI ESlRm
Fig. 1, Total HDL cholesterol
lm3lllfll
lmllLEu
DOSE
and HDL cholesterol
subfractions
by estrogen
dose.
36
TABLE
4
HDL-CHOLESTEROL (mg/dl) LEVELS ADJUSTED FOR BMI, ALCOHOL INTAKE, PHYSICAL ACTIVITY AND CIGARETTE SMOKING - ESTROGEN AND NON-ESTROGEN POSTMENOPAUSAL WOMEN
HDL-cholesterol HDL,-cholesterol HDL,-cholesterol
** **
Estrogen (n = 48)
Non-estrogen (n = 246)
14.2 35.4 38.7
61.8 23.2 38.6
** P < 0.01 consumption, body mass index, physical activity and cigarette smoking. As presented, the adjusted data differed very little from the unadjusted data. The increase in HDL-total cholesterol and HDL,-C in the estrogen users remained. Again, there was no significant difference between the estrogen and non-estrogen users for HDL,. These results suggest that estrogen use is independently related to HDL-total and HDL,-C levels. In order to determine whether the increased HDL-C concentrations in the estrogen users were associated with liver function, particularly hepatic microsomal activity, we analyzed the concentrations of the liver enzymes, serum glutamic oxalacetic transaminase (SGOT), y-glutamyl transpeptidase (GGTP) and glutamate pyruvic transaminase (SGPT). As presented in Table 5, the SGOT and SGPT concentrations were significantly lower in the estrogen group; no difference was observed in GGTP levels. This finding is opposite to the hypothesis that the increased HDL-C concentrations in estrogen users reflected an estrogen stimulation of hepatic microsomal activity. Instead, estrogen may act as an enzyme inhibitor as determined by the decreased liver enzymes in the estrogen users. However, of the three enzymes, GGTP is the most sensitive to liver function. Since no significant differences in this enzyme were observed, it may be premature to suggest that estrogen is an enzyme inhibitor.
TABLE
5
LIVER MEN
FUNCTION
TESTS
IN ESTROGEN
AND
NON-ESTROGEN
POSTMENOPAUSAL
Non-estrogen (n = 246)
Estrogen (n = 48) x
(SD?)
x
(SD%)
SGF’T (U/l) * SGOT (U/l) * GGTP (U/l)
8.8 19.4 26.1
(7.4) (8.2) (21.1)
12.3 25.1 26.5
(4.9) (9.8) (18.1)
* P < 0.05. x = mean; SD? = standard
deviation.
WO-
37
Discussion The results of the research clearly indicate that postmenopausal estrogen users have higher HDL-C levels than non-users. Exogenous estrogen use appears to be an important determinant of HDL-C. Additionally, the locus of the increase is in the HDL, subfraction. The primary relationship to HDL, is important since HDL, may be the subfraction more closely linked to cardiovascular risk [5]; however, little is known about the relationship of HDL, to risk of heart attack in women. The relationship between total cholesterol, triglycerides and exogenous estrogen is not clear. Some studies have found that estrogen raises the triglyceride concentration and lowers total cholesterol [10,21]. However, other studies have found no relationship to triglycerides or cholesterol [5]. The effect on triglycerides may be determined by the type of estrogen used. Synthetic estrogens may increase the triglyceride levels [21-331, but natural estrogens may not have this effect [6]. It is clear, however, that the strongest relationship of estrogen to lipoproteins is to HDL-C, in particular HDL,. The mechanism by which exogenous estrogen raises HDL-C concentrations does not appear to be through alterations in hepatic function. Although estrogen use was related to the liver function, it was in the opposite direction to that which was expected. Cowan et al. [24] have also reported lower levels of SGOT in estrogen users. It is not clear why estrogen administration would result in lower levels of liver enzymes; however, the assessment of hepatic function by liver enzymes is relatively crude and perhaps, more indepth assessments of liver function need to be made. It has been suggested that increased levels of liver enzymes may be present at the initiation of hormone therapy, but, with long-term use of the hormone, an adaptation in the liver may occur [24,25]. This adaptation may result in lower levels of enzymes. We have found similar results with endogenous testosterone where serum testosterone levels were significantly positively related to HDL-C levels, independent of liver enzymes [26]. Coupled with the current results, it appears that sex hormones may raise HDL-C concentrations through an alternative, unknown mechanism, different from that of alcohol, phenobarbital, clofibrate or physical activity. One possible mechanism may be an effect on hepatic endothelial lipase activity. Preliminary evidence has suggested that hepatic lipase may contribute to the catabolism of HDL [27]. Some researchers have suggested that the whole HDL, particle may not be removed. HDL, may lose surface phospholipids, followed by a release of cholesterol esters to the liver. HDL, may be subsequently returned to the circulation [28,29]. Since hepatic lipase appears to be sensitive to estrogens [9], HDL, may be affected, while HDL, may not. Estrogen users may also have higher HDL-C and HDL,-C concentrations through an effect on the lipoprotein lipase enzyme system. Lipoprotein lipase is the key enzyme in the catabolism of triglyceride-rich lipoproteins. It is not known to what extent variations of plasma HDL can be accounted for by the reciprocal relationship with the metabolism of triglyceride-rich lipoproteins. Insulin has been shown to stimulate lipoprotein lipase activity [30]. Adolescent and young women were noted
38
to have a greater insulin response to a glucose challenge than men of similar ages [31]. In addition, women on exogenous hormones had lower fasting glucose levels [32]. Therefore, estrogen administration may either stimulate insulin production or decrease insulin resistance which in turn, may stimulate lipoprotein lipase activity, resulting in a rise in HDL-C and a fall in triglyceride concentrations. However, the decrease in insulin resistance may only be secondary to the lower adiposity observed in estrogen users. We are currently investigating the associations between fasting insulin concentrations, HDL-C levels and hormone use. Further research is also needed to examine the interrelationships of estrogen use, hepatic function, hepatic endothelial lipase and HDL-C. References 1 Bain, C., Willett, W., Hennekens, C.H., Rosner, B., Belanger, C. and Speizer, F.E.. Use of postmenopausal hormones and risk of myocardial infarction, Circulation, 64 (1982) 42. 2 Pfeffer, R.I., Whipple, G.H., Kurosaki, T.T. and Chapman, J.M., Coronary risk and estrogen use in postmenopausal women, Amer. J. Epidemiol., 107 (1978) 479. 3 Bush, T.L., Cowan, L.D., Barrett-Connor, E.B., Criqui, M.H., Karon, J.M., Wallace, R.B., Tyroler, A. and Rifkind, B.M., Estrogen use and all cause mortality - Preliminary results from the Lipid Research Clinics Program Follow-up Study, J. Amer. Med. Ass., 249 (1983) 903. 4 Gordon, T., Castelli, W.P., Hjortland, M.C., Kannel, W.B. and Dawber, T.R., High density lipoprotein as a protective factor against coronary heart disease - The Framingham Study, Amer. J. Med., 62 (1977) 707. 5 Ballantyne, F.C., Clark, R.S., Simpson, H.S. and Ballantyne, D., High density and low density lipoprotein subfractions in survivors of myocardial infarction and in control subjects, Metabolism, 31 (1982) 433. 6 Walletin, L. and Larsson-Cohn, U., Metabolic and hormonal effects of postmenopausal oestrogen replacement treatment, Acta Endocrinol., 86 (1977) 597. 7 Bradley, D.D., Wingerd, J., Petitti, D.B., Krauss, R.M. and Ramcharan, S., Serum high density lipoprotein cholesterol in women using oral contraceptives, estrogens and progestins, N. Eng. J Med., 299 (1978) 17. 8 Krauss, R.M., Lindgren, F.T., Wingerd, J., Bradley, D.D. and Ramcharan, S., Effects of estrogen and progestins on high density lipoproteins, Lipids, 14 (1979) 113. 9 Tikkanen, M.J., Nikkila, E.A., Kuusi, T. and Sipinem, S., High density lipoprotein, and hepatic lipase - Reciprocal changes produced by estrogen and norgestrel, J. Clin. Endocrinol. Med., 54 (1982) 1113. 10 Barrett-Connor, E., Brown, V., Turner, J., Austin, M. and Criqui, M.H., Heart disease risk factors and hormone use in postmenopausal women, J. Amer. Med. Ass., 241 (1979) 2167. 11 Tikkanen, M., Kuusi, T., Vartianen, E. and Nikkila, E.A., Treatment of postmenopausal hypercholesteremia with estradiol, Acta Obst. Gynecol. Stand., Suppl. 88 (1972) 83. 12 LaPorte, R.E., Falvo-Gerard, L., Kuller, L., Dai, W., Bates, M., Cresanta, J., Williams, K. and Palkin, D., The relationship between alcohol consumption, liver enzymes and high density lipoprotein cholesterol, Circulation, 64 (1981) 111-67. 13 Heller, R.F., Coronary heart disease, cancer, lipoproteins, and the effects of clofibrate - Is enzyme induction a common link and are lipoproteins red herrings? Lancet, ii (1981) 1258. 14 Kuller, L.H., Hulley, S.B., LaPorte, R.E., Neaton, J. and Dai, W.S., Environmental determinants, liver function, and high density lipoprotein cholesterol levels, Amer. J. Epidemiol., 117 (1983) 4060. 15 Barrett-Connor, E. and Suarez, L., A community study of alcohol and other factors associated with the distribution of high density lipoprotein cholesterol in older vs younger men, Amer. J. Epidemiol., 115 (1982) 888. 16 Bolton, C.H., Jackson, L. and Roberts, C.J., Enzyme induction and serum and lipoprotein lipids - A study of glutethimide in normal subjects, Clin. Sci., 58 (1980) 419.
39
17 Pynadath, T.T. and Chanapai, S., Elevation of serum HDL and HDL cholesterol in cholesterol-fed male rabbits treated with estrogen, Atherosclerosis, 28 (1981) 255. 18 Paffenbarger, R.S., Wing, A.L. and Hyde, R., Physical activity as an index of heart attack risk in college alumni, Amer. J. Epidemiol., 108 (1978) 161. 19 Gidez, L.I., Miller, G.J., Burstein, M. and Eder, H.A., Analysis of HDL subclasses by a precipitation procedure - Correlation with preparative and analytical ultracentrifugation in Report of the HDL Methodology Workshop (NIH Publication 79-1661), National Institutes of Health, Bethesda, MD, 1977. 20 Allain, C.C., Poon, L.S., Chan, C.S.G., Richmond, W. and Fu, P.C., Enzymatic determination of total serum cholesterol, Clin. Chem., 20 (1974) 470. 21 Tikkanen, M.J., Nikkila, E.A. and Vartiainen, E., Natural oestrogen as an effective treatment for Type II hyperlipoproteinaemia in postmenopausal women, Lancet, ii (1978) 490. 22 Bolton, C., Ellwood, M., Hartog, M., Martin, R., Rowe, A.S. and Winsley, R.T., Comparisons of the effects of ethinyl estradiol and conjugated equine estrogens in oophorectomized women, Clin. Endocr., 4 (1975) 131. 23 Buckman, M., Johnson, T., Ellis, H., Srivastava, L. and Peake, C.T., Differential lipemic and proteinemic response to oral ethinyl estradiol and parenteral estradiol cypionate, Metabolism, 29 (1980) 803. 24 Cowan, L.D., Wallace, R.B., Barrett-Connor, E., Hunninghake, D., Pomrehn, P., Whal, P. and Heiss, C., Alterations in clinical chemistry measures associated with oral contraceptive and estrogen use - The Lipid Research Clinics Program Prevalence Study, J. Reprod. Med.. 27 (1982) 275. 25 Swyer, G.I.M. and Little, V., Absence of hepatic impairment in long term oral contraceptive users, Brit. Med. J., I (1965) 1412. 26 Gutai, J., LaPorte, R.E., Kuller, L.H., Dai, W., Falvo-Gerard, L. and Caggiula, A., Plasma testosterone, high density lipoprotein cholesterol and other lipoproteins, Amer. J. Cardiol., 48 (1981) 897. 27 Nikkila, E.A., Kuusi, T., Harno, K., Tikkanen, M. and Taskinen, M.R., Lipoprotein lipase and hepatic endothelial lipase are key enzymes in the metabolism of plasma high density lipoproteins, particularly HDL,. In: A.M. Gotto, L.C. Smith and B. Allen (Eds.), International Symposium on Atherosclerosis, 5th edition, Houston, TX, 1980, p. 387. 28 Nikkila, E.A., Kuusi, T. and Taskinen. M., Role of lipoprotein lipase and hepatic lipase in the metabolism of high density lipoproteins - A novel concept on cholesterol transport in HDL cycle. In: L.A. Carlson and B. Pernow (Eds.). Metabolic Risk Factors in Ischemic Cardiovascular Disease, Raven Press, New York, 1982, p. 205. 29 Kuusi, T.. Sarrinen, P. and Nikkila, E.A., Evidence for the mode of hepatic endothelial lipase in the metabolism of plasma high density lipoproteins in man, Atherosclerosis, 36 (1980) 589. 30 Nil&la, E.A., Metabolic and endocrine control of plasma high density lipoprotein concentration. In: A.M. Gotto, N.E. Miller and M.F. Oliver (Eds.), High Density Lipoprotein and Atherosclerosis, Elsevier/North-Holland Biomedical Press, Amsterdam, 1978, p. 117. 3 I Orchard, T.J., Becker, D.J., Kuller, L.H., Wagener, D.K., LaPorte, R.E. and Drash, A.L., Age and sex variations in glucose tolerance and insulin responses - Parallels with cardiovascular risk, J. Chron. Dis., 35 (1982) 123. 32 Barrett-Conner, E., Factors associated with the distribution of fasting plasma glucose in an adult community, Amer. J Epidemiol., I 12 (1980) 5 18.
31 Ltd.
ATH 03391
Menopausal Estrogen Use, High Density Lipoprotein Cholesterol Subfractions and Liver Function Jane A. Cauley
‘, Ronald
E. LaPorte ‘, Lewis H. Kuller ‘, Margaret and Rivka Black Sandler *
Bates ’
’Department of Epidemiology, Graduate School of Public Health and ’School of Health Related Professions,
University of Pittsburgh,
Pittsburgh, PA 15261 (U.S.A.)
(Received 13 August, 1982) (Revised, received 27 April, 1983) (Accepted 28 April, 1983)
Summary
Forty-eight menopausal women taking exogenous estrogen were compared with 246 postmenopausal women not on estrogen. The estrogen users had significantly higher total high density lipoprotein (HDL) (76.0 vs 61.4 mg/dl) and HDL, (36.7 vs 23.0 mg/dl) cholesterol than the controls. There was a similar concentration of HDL, cholesterol for the two groups (39.2 for the estrogen users and 38.4 for the controls). A dose-response was evident between the amount of daily estrogen and HDL-total and HDL, cholesterol. The significant differences between the two groups remained after adjusting for body composition, alcohol intake, cigarette smoking and physical activity. There was a significant difference between the two groups in liver function as measured by the liver enzymes, SGOT, SGPT, with liver enzyme concentrations lower in the estrogen users. The results indicate that the increase in the total HDL cholesterol as a result of menopausal estrogen is primarily the result of increased HDL,. The increase could not be explained by alterations in hepatic microsomal activity as measured by liver enzymes since estrogen users had lower concentrations of liver enzymes than non-estrogen users. Key words:
Estrogen treatment - High density zymes - Liver function - Menopause
lipoprotein
cholesterol - Liver
en-
This research was supported in part by a grant from the Alcohol Beverage Medical Research Foundation and Grants AM-l 180 and AM-21 190 from the National Institutes of Health. Reprint requests to Dr. Cauley, Department of Epidemiology, University of Pittsburgh, Pittsburgh, PA 15261, U.S.A.
0021-9150/83/$03.00
0 1983 Elsevier Scientific
Publishers
Ireland,
Ltd.
32
Introduction The relationship of postmenopausal estrogen use to cardiovascular disease is controversial. One report has suggested that postmenopausal estrogen use may be protective against cardiovascular disease [l] whereas others have demonstrated no relationship [2]. Recently, preliminary results from the Lipid Research Clinics (LRC) have suggested an inverse association between estrogen use and total mortality [3]. Yet, the mechanism by which exogenous estrogen can affect risk is not known, but may be through alterations in lipids, in particular the level of high density lipoprotein cholesterol (HDL-C). High density lipoprotein cholesterol is a primary predictor of coronary heart disease (CHD) risk among postmenopausal women [4]. Recently, the heterogeneous nature of the HDL has been recognized, and it has been suggested that the subfraction HDL, cholesterol (HDL,-C) may be a more effective epidemiologic discriminant of CHD risk than the total HDL-C [5]. The research examining the effect of estrogen use on the total HDL-C has consistently reported that estrogen users had higher levels of total HDL-C [6,7]. The effect of postmenopausal estrogen use on the HDL-C subfractions is less clear. Little is known concerning the relationship of estrogen use to HDL-C subfractions. Krauss et al. [8] noted significantly higher levels of HDL, in 11 women, aged 44-66, taking natural estrogens. In support of this, Tikkanen et al. [9] described a 20% increase in the HDL, cholesterol during oral estradiol treatment in 16 women. The failure to control for other known HDL-C determinants such as the degree of obesity, alcohol consumption, physical activity and cigarette smoking, may have influenced their results since there has been some suggestion that estrogen users may differ from non-users on other characteristics other than their lipid values. In the LRC, women on exogenous hormones were found to be thinner [lo]. In addition to failing to control for confounding variables, the sample sizes in both reports were quite small. The current research was designed to investigate the relationship of postmenopausal estrogen utilization to lipoprotein cholesterol concentrations including HDL, and HDL,. In addition, the research controlled for other known determinants of HDL-C in order to evaluate the independent relationship of menopausal estrogen use to the HDL-C subfractions. The mechanisms by which estrogen use is related to increased HDL-C levels have not been fully elucidated. Exogenous estrogen may increase HDL-C levels through an effect on hepatic lipase [ 111. Estrogen may also increase HDL-C through enzyme induction. We, as well as others, have suggested that several determinants of HDL-C, including alcohol consumption, drugs such as phenobarbital, glutethimide, clofibrate, and perhaps physical activity may raise HDL-C levels through increased hepatic microsomal activity [ 12-161. Pynadath and Chanapai have suggested that estrogen has a similar chemical structure as known enzyme inducers [ 171, and thus conceivably could raise HDL-C by the same mechanism as alcohol. We, therefore, examined liver enzymes as indirect measures of liver function in order to determine if the relationship of exogenous estrogen use to HDL-C was possibly the result of increased hepatic microsomal activity or alteration of liver function.
33
Methods Two hundred and ninety-four postmenopausal women, 246 non-estrogen users and 48 estrogen users were recruited as an ancillary study to research investigating the epidemiology of bone loss in postmenopausal women. As part of the research, blood samples were taken for lipoprotein cholesterol analysis and liver enzymes. Upon entry into the study, all women visited the clinic and completed a medical history questionnaire, including current and past hormone use, the Paffenbarger questionnaire for assessment of physical activity which converts activity above and [ 181 and a cigarette and alcohol beyond activities of daily living to kcal/wk questionnaire. Anthropometric measurements were taken, including height, weight, grip strength and the triceps skinfold. The lipid analyses were performed by the Heinz Nutrition Laboratory at the Graduate School of Public Health, University of Pittsburgh. The laboratory has received certification of quality control from the Center for Disease Control. High density lipoprotein cholesterol concentrations were determined by precipitation procedures using heparin and manganese chloride. The subfraction HDL,-C was then precipitated by dextran sulfate. The soluble cholesterol represented HDL,-C [ 191. Cholesterol analyses were performed using the enzymatic method [20]. Results Table 1 presents the baseline characteristics comparing the estrogen users with the non-estrogen users. Both groups were of similar age and menopausal status. The estrogen users had a lower body mass index (BMI) [wt (kg)/ht*(m)], and had higher grip strength. In the estrogen group, all women reported the use of conjugated equine estrogens, but were on different dosages and regimes. Because of the various drug regimes, the dose was summed over a l-month period, and then divided by 30 to estimate a daily estrogen dose. The frequency distribution of daily estrogen dose is presented in Table 2. The majority of the women were on relatively low doses of estrogen. Alcohol consumption was estimated from questionnaire data which asked the frequency and type of beverage consumed. This was converted to ounces of ethanol. Information on cigarette smoking was also obtained from the questionnaire. There were 12 or 23.1% cigarette smokers in the estrogen group and 30 or 12.4% in the non-estrogen group. The lipid profiles for the two groups are presented in Table 3. There was little difference in the total cholesterol and the triglyceride levels. The total HDL and the HDL,-C concentrations were both significantly higher in the estrogen users. The largest difference was in the HDL, subfraction where HDL,-C was 13.8 mg/dl (38%) higher in the estrogen users. There was little difference between the two groups for the HDL, subfraction. The data were also examined for an estrogen dose effect on the HDL-C subfractions. The daily estrogen dose was positively related to the total HDL-C (r = 0.35, P -c O.OOl), and to the HDL,-C (r = 0.41, P < 0.001) in the univariate
34 TABLE
1
BASELINE
CHARACTERISTICS
OF 151 POSTMENOPAUSAL Estrogen users (n = 48)
Non-estrogen (n = 246)
x
(SD_?) 56.6 8.8 47.5
Age (yrs) Years menopausal Age at menopause
1299.5 25.2
(4.0) (23.7)
145.2 63.1 25.2 27.1
(2.8) (3.6) (4.8)
Alcohol consumption (ounces ethanol/day)
0.44
(0.49)
0.36
Daily estrogen
0.37
(0.19)
0.00
dose (mg)
* P < 0.05. X = mean; SD,? = standard
users
(SD.?) 51.6 8.5 49.3
(1027.0)
139.4 64.3 23.1 26.3
Weight (lbs) Height (in) BMI (kg/m*) * Triceps skinfold (mm)
x
(5.3) (5.2) (5.5)
1441.1 21.6
kcal/week (Paffenbarger) Grip strength (kg) *
WOMEN
(4.2) (5.8) (4.6) (1006.7) (4.8) (25.0) (2.2) (3.9) (5.5)
(0.59)
deviation.
analysis. There was no relationship between estrogen dose and HDL,-C, total cholesterol or triglycerides. Since body weight may influence the pharmacologic effect of the drug, the relationship between dose per pound of body weight and HDL-C subfractions was also examined. In this analysis, the estrogen dose per pound continued to be positively related to the total HDL-C (r = 0.28, P < 0.03)
TABLE
2
FREQUENCY MEN Daily conjugated estrogen (mg)
DISTRIBUTION
equine
OF DAILY
ESTROGEN
n
< 0.10 0.1 l-0.20 0.21-0.30 0.3 l-0.40 0.41-0.50 0.5 l-0.60 0.6 l-0.70 0.7 l-0.80 0.8 l-0.90
2 6 13 5 14 3 3 0 0
0.91-1.00
2 4%
DOSE
IN 48 POSTMENOPAUSAL
(%)
(4.2) (12.5) (27.1) (10.4) (29.2) (6.25) (6.25) _ _ (4.2) (loorc]
WO-
35
TABLE LIPID
3 PROFILES
(mg/dl)
-
ESTROGEN
AND NON-ESTROGEN
Triglycerides Total cholesterol HDL-cholesterol HDL,-cholesterol
** **
HDL,-cholesterol
**p < 0.01. X = mean; SD.%?= standard
WOMEN
Non-estrogen (n = 246)
Estrogen (n = 48)
Lipid
POSTMENOPAUSAL
7
(SD?)
x
(SDY)
110.4 229.4 76.0 36.7
(47.0) (35.4) (17.1) (15.6)
113.4 228.0 61.4 23.0
(59.5) (41.2) (14.5) (11.0)
39.2
(7.0)
38.4
(7.4)
deviation.
and to HDL,-C (r = 0.28, P < 0.03), although the magnitude of the correlation coefficient decreased. These findings are graphically presented in Fig. 1 where HDL-total cholesterol and HDL,- and HDL,-C are examined by estrogen dose. Tertile I corresponds to a daily estrogen dosage of 0.04-0.25 mg; Tertile II; 0.26-0.46 mg; Tertile III: 0.47-0.96 mg. A linear relationship is present for both the total HDL-C and the HDL,-C. Since the estrogen users were thinner and there was a tendency for the estrogen users to consume more alcohol and be more active, these variables may have contributed to the increased HDL-total cholesterol and the HDL,-C levels in the estrogen users. Therefore, analysis of co-variance was employed to adjust for potential confounders. Table 4 presents the mean HDL-C levels adjusted for alcohol mo-
/------A
la-
3
a
-5 d 3 J
so-
40-
I
-_______-------& /xX-
/
//
lsged
A BDL-1 C+tOL
0
x HDL-2 CHOL ---0 HDL-3 -__-A- c+iOL
NOBl’ROGD(
lmnl.EI ESlRm
Fig. 1, Total HDL cholesterol
lm3lllfll
lmllLEu
DOSE
and HDL cholesterol
subfractions
by estrogen
dose.
36
TABLE
4
HDL-CHOLESTEROL (mg/dl) LEVELS ADJUSTED FOR BMI, ALCOHOL INTAKE, PHYSICAL ACTIVITY AND CIGARETTE SMOKING - ESTROGEN AND NON-ESTROGEN POSTMENOPAUSAL WOMEN
HDL-cholesterol HDL,-cholesterol HDL,-cholesterol
** **
Estrogen (n = 48)
Non-estrogen (n = 246)
14.2 35.4 38.7
61.8 23.2 38.6
** P < 0.01 consumption, body mass index, physical activity and cigarette smoking. As presented, the adjusted data differed very little from the unadjusted data. The increase in HDL-total cholesterol and HDL,-C in the estrogen users remained. Again, there was no significant difference between the estrogen and non-estrogen users for HDL,. These results suggest that estrogen use is independently related to HDL-total and HDL,-C levels. In order to determine whether the increased HDL-C concentrations in the estrogen users were associated with liver function, particularly hepatic microsomal activity, we analyzed the concentrations of the liver enzymes, serum glutamic oxalacetic transaminase (SGOT), y-glutamyl transpeptidase (GGTP) and glutamate pyruvic transaminase (SGPT). As presented in Table 5, the SGOT and SGPT concentrations were significantly lower in the estrogen group; no difference was observed in GGTP levels. This finding is opposite to the hypothesis that the increased HDL-C concentrations in estrogen users reflected an estrogen stimulation of hepatic microsomal activity. Instead, estrogen may act as an enzyme inhibitor as determined by the decreased liver enzymes in the estrogen users. However, of the three enzymes, GGTP is the most sensitive to liver function. Since no significant differences in this enzyme were observed, it may be premature to suggest that estrogen is an enzyme inhibitor.
TABLE
5
LIVER MEN
FUNCTION
TESTS
IN ESTROGEN
AND
NON-ESTROGEN
POSTMENOPAUSAL
Non-estrogen (n = 246)
Estrogen (n = 48) x
(SD?)
x
(SD%)
SGF’T (U/l) * SGOT (U/l) * GGTP (U/l)
8.8 19.4 26.1
(7.4) (8.2) (21.1)
12.3 25.1 26.5
(4.9) (9.8) (18.1)
* P < 0.05. x = mean; SD? = standard
deviation.
WO-
37
Discussion The results of the research clearly indicate that postmenopausal estrogen users have higher HDL-C levels than non-users. Exogenous estrogen use appears to be an important determinant of HDL-C. Additionally, the locus of the increase is in the HDL, subfraction. The primary relationship to HDL, is important since HDL, may be the subfraction more closely linked to cardiovascular risk [5]; however, little is known about the relationship of HDL, to risk of heart attack in women. The relationship between total cholesterol, triglycerides and exogenous estrogen is not clear. Some studies have found that estrogen raises the triglyceride concentration and lowers total cholesterol [10,21]. However, other studies have found no relationship to triglycerides or cholesterol [5]. The effect on triglycerides may be determined by the type of estrogen used. Synthetic estrogens may increase the triglyceride levels [21-331, but natural estrogens may not have this effect [6]. It is clear, however, that the strongest relationship of estrogen to lipoproteins is to HDL-C, in particular HDL,. The mechanism by which exogenous estrogen raises HDL-C concentrations does not appear to be through alterations in hepatic function. Although estrogen use was related to the liver function, it was in the opposite direction to that which was expected. Cowan et al. [24] have also reported lower levels of SGOT in estrogen users. It is not clear why estrogen administration would result in lower levels of liver enzymes; however, the assessment of hepatic function by liver enzymes is relatively crude and perhaps, more indepth assessments of liver function need to be made. It has been suggested that increased levels of liver enzymes may be present at the initiation of hormone therapy, but, with long-term use of the hormone, an adaptation in the liver may occur [24,25]. This adaptation may result in lower levels of enzymes. We have found similar results with endogenous testosterone where serum testosterone levels were significantly positively related to HDL-C levels, independent of liver enzymes [26]. Coupled with the current results, it appears that sex hormones may raise HDL-C concentrations through an alternative, unknown mechanism, different from that of alcohol, phenobarbital, clofibrate or physical activity. One possible mechanism may be an effect on hepatic endothelial lipase activity. Preliminary evidence has suggested that hepatic lipase may contribute to the catabolism of HDL [27]. Some researchers have suggested that the whole HDL, particle may not be removed. HDL, may lose surface phospholipids, followed by a release of cholesterol esters to the liver. HDL, may be subsequently returned to the circulation [28,29]. Since hepatic lipase appears to be sensitive to estrogens [9], HDL, may be affected, while HDL, may not. Estrogen users may also have higher HDL-C and HDL,-C concentrations through an effect on the lipoprotein lipase enzyme system. Lipoprotein lipase is the key enzyme in the catabolism of triglyceride-rich lipoproteins. It is not known to what extent variations of plasma HDL can be accounted for by the reciprocal relationship with the metabolism of triglyceride-rich lipoproteins. Insulin has been shown to stimulate lipoprotein lipase activity [30]. Adolescent and young women were noted
38
to have a greater insulin response to a glucose challenge than men of similar ages [31]. In addition, women on exogenous hormones had lower fasting glucose levels [32]. Therefore, estrogen administration may either stimulate insulin production or decrease insulin resistance which in turn, may stimulate lipoprotein lipase activity, resulting in a rise in HDL-C and a fall in triglyceride concentrations. However, the decrease in insulin resistance may only be secondary to the lower adiposity observed in estrogen users. We are currently investigating the associations between fasting insulin concentrations, HDL-C levels and hormone use. Further research is also needed to examine the interrelationships of estrogen use, hepatic function, hepatic endothelial lipase and HDL-C. References 1 Bain, C., Willett, W., Hennekens, C.H., Rosner, B., Belanger, C. and Speizer, F.E.. Use of postmenopausal hormones and risk of myocardial infarction, Circulation, 64 (1982) 42. 2 Pfeffer, R.I., Whipple, G.H., Kurosaki, T.T. and Chapman, J.M., Coronary risk and estrogen use in postmenopausal women, Amer. J. Epidemiol., 107 (1978) 479. 3 Bush, T.L., Cowan, L.D., Barrett-Connor, E.B., Criqui, M.H., Karon, J.M., Wallace, R.B., Tyroler, A. and Rifkind, B.M., Estrogen use and all cause mortality - Preliminary results from the Lipid Research Clinics Program Follow-up Study, J. Amer. Med. Ass., 249 (1983) 903. 4 Gordon, T., Castelli, W.P., Hjortland, M.C., Kannel, W.B. and Dawber, T.R., High density lipoprotein as a protective factor against coronary heart disease - The Framingham Study, Amer. J. Med., 62 (1977) 707. 5 Ballantyne, F.C., Clark, R.S., Simpson, H.S. and Ballantyne, D., High density and low density lipoprotein subfractions in survivors of myocardial infarction and in control subjects, Metabolism, 31 (1982) 433. 6 Walletin, L. and Larsson-Cohn, U., Metabolic and hormonal effects of postmenopausal oestrogen replacement treatment, Acta Endocrinol., 86 (1977) 597. 7 Bradley, D.D., Wingerd, J., Petitti, D.B., Krauss, R.M. and Ramcharan, S., Serum high density lipoprotein cholesterol in women using oral contraceptives, estrogens and progestins, N. Eng. J Med., 299 (1978) 17. 8 Krauss, R.M., Lindgren, F.T., Wingerd, J., Bradley, D.D. and Ramcharan, S., Effects of estrogen and progestins on high density lipoproteins, Lipids, 14 (1979) 113. 9 Tikkanen, M.J., Nikkila, E.A., Kuusi, T. and Sipinem, S., High density lipoprotein, and hepatic lipase - Reciprocal changes produced by estrogen and norgestrel, J. Clin. Endocrinol. Med., 54 (1982) 1113. 10 Barrett-Connor, E., Brown, V., Turner, J., Austin, M. and Criqui, M.H., Heart disease risk factors and hormone use in postmenopausal women, J. Amer. Med. Ass., 241 (1979) 2167. 11 Tikkanen, M., Kuusi, T., Vartianen, E. and Nikkila, E.A., Treatment of postmenopausal hypercholesteremia with estradiol, Acta Obst. Gynecol. Stand., Suppl. 88 (1972) 83. 12 LaPorte, R.E., Falvo-Gerard, L., Kuller, L., Dai, W., Bates, M., Cresanta, J., Williams, K. and Palkin, D., The relationship between alcohol consumption, liver enzymes and high density lipoprotein cholesterol, Circulation, 64 (1981) 111-67. 13 Heller, R.F., Coronary heart disease, cancer, lipoproteins, and the effects of clofibrate - Is enzyme induction a common link and are lipoproteins red herrings? Lancet, ii (1981) 1258. 14 Kuller, L.H., Hulley, S.B., LaPorte, R.E., Neaton, J. and Dai, W.S., Environmental determinants, liver function, and high density lipoprotein cholesterol levels, Amer. J. Epidemiol., 117 (1983) 4060. 15 Barrett-Connor, E. and Suarez, L., A community study of alcohol and other factors associated with the distribution of high density lipoprotein cholesterol in older vs younger men, Amer. J. Epidemiol., 115 (1982) 888. 16 Bolton, C.H., Jackson, L. and Roberts, C.J., Enzyme induction and serum and lipoprotein lipids - A study of glutethimide in normal subjects, Clin. Sci., 58 (1980) 419.
39
17 Pynadath, T.T. and Chanapai, S., Elevation of serum HDL and HDL cholesterol in cholesterol-fed male rabbits treated with estrogen, Atherosclerosis, 28 (1981) 255. 18 Paffenbarger, R.S., Wing, A.L. and Hyde, R., Physical activity as an index of heart attack risk in college alumni, Amer. J. Epidemiol., 108 (1978) 161. 19 Gidez, L.I., Miller, G.J., Burstein, M. and Eder, H.A., Analysis of HDL subclasses by a precipitation procedure - Correlation with preparative and analytical ultracentrifugation in Report of the HDL Methodology Workshop (NIH Publication 79-1661), National Institutes of Health, Bethesda, MD, 1977. 20 Allain, C.C., Poon, L.S., Chan, C.S.G., Richmond, W. and Fu, P.C., Enzymatic determination of total serum cholesterol, Clin. Chem., 20 (1974) 470. 21 Tikkanen, M.J., Nikkila, E.A. and Vartiainen, E., Natural oestrogen as an effective treatment for Type II hyperlipoproteinaemia in postmenopausal women, Lancet, ii (1978) 490. 22 Bolton, C., Ellwood, M., Hartog, M., Martin, R., Rowe, A.S. and Winsley, R.T., Comparisons of the effects of ethinyl estradiol and conjugated equine estrogens in oophorectomized women, Clin. Endocr., 4 (1975) 131. 23 Buckman, M., Johnson, T., Ellis, H., Srivastava, L. and Peake, C.T., Differential lipemic and proteinemic response to oral ethinyl estradiol and parenteral estradiol cypionate, Metabolism, 29 (1980) 803. 24 Cowan, L.D., Wallace, R.B., Barrett-Connor, E., Hunninghake, D., Pomrehn, P., Whal, P. and Heiss, C., Alterations in clinical chemistry measures associated with oral contraceptive and estrogen use - The Lipid Research Clinics Program Prevalence Study, J. Reprod. Med.. 27 (1982) 275. 25 Swyer, G.I.M. and Little, V., Absence of hepatic impairment in long term oral contraceptive users, Brit. Med. J., I (1965) 1412. 26 Gutai, J., LaPorte, R.E., Kuller, L.H., Dai, W., Falvo-Gerard, L. and Caggiula, A., Plasma testosterone, high density lipoprotein cholesterol and other lipoproteins, Amer. J. Cardiol., 48 (1981) 897. 27 Nikkila, E.A., Kuusi, T., Harno, K., Tikkanen, M. and Taskinen, M.R., Lipoprotein lipase and hepatic endothelial lipase are key enzymes in the metabolism of plasma high density lipoproteins, particularly HDL,. In: A.M. Gotto, L.C. Smith and B. Allen (Eds.), International Symposium on Atherosclerosis, 5th edition, Houston, TX, 1980, p. 387. 28 Nikkila, E.A., Kuusi, T. and Taskinen. M., Role of lipoprotein lipase and hepatic lipase in the metabolism of high density lipoproteins - A novel concept on cholesterol transport in HDL cycle. In: L.A. Carlson and B. Pernow (Eds.). Metabolic Risk Factors in Ischemic Cardiovascular Disease, Raven Press, New York, 1982, p. 205. 29 Kuusi, T.. Sarrinen, P. and Nikkila, E.A., Evidence for the mode of hepatic endothelial lipase in the metabolism of plasma high density lipoproteins in man, Atherosclerosis, 36 (1980) 589. 30 Nil&la, E.A., Metabolic and endocrine control of plasma high density lipoprotein concentration. In: A.M. Gotto, N.E. Miller and M.F. Oliver (Eds.), High Density Lipoprotein and Atherosclerosis, Elsevier/North-Holland Biomedical Press, Amsterdam, 1978, p. 117. 3 I Orchard, T.J., Becker, D.J., Kuller, L.H., Wagener, D.K., LaPorte, R.E. and Drash, A.L., Age and sex variations in glucose tolerance and insulin responses - Parallels with cardiovascular risk, J. Chron. Dis., 35 (1982) 123. 32 Barrett-Conner, E., Factors associated with the distribution of fasting plasma glucose in an adult community, Amer. J Epidemiol., I 12 (1980) 5 18.