0021 -972X/88/6602-0314$02.00/0Journal of Clinical Endocrinology and MetabolismCopyright © 1988 by The Endocrine Society

Vol. 66, No. 2Printed in U.S.A.

Changes in Plasma Lipoproteins During VariousAndrogen Suppression Therapies in Men With ProstaticCarcinoma: Effects of Orchiectomy, Estrogen, andCombination Treatment With Luteinizing Hormone-Releasing Hormone Agonist and FlutamideSITAL MOORJANI, ANDRE DUPONT, FERNAND LABRIE, PAUL-J. LUPIEN,CLAUDE GAGNE, DANIEL BRUN, MICHEL GIGUERE, ALAIN BELANGER, ANDLIONELLO CUSAN

Lipid Research Unit (S.M., P.-J.L., C.G., D.B.) and the Molecular Endocrinology Group (A.D., F.L., M.G.,A.B., L.C.), Laval University Hospital Research Center, and the Departments of Biochemistry, Physiology,and Medicine, Faculty of Medicine, Laval University, Quebec City, Quebec, Canada

ABSTRACT. Cardiovascular complications are a well recog-nized side-effect of antihormonal therapy in men with prostaticcarcinoma. We studied changes in plasma lipoproteins in pa-tients with prostate cancer during treatment with several andro-gen suppression therapies. Estrogen, orchiectomy, and a com-bination of LHRH agonist and antiandrogen (flutamide) reducedplasma testosterone concentrations (89-92%) and plasma estra-diol decreased by 85%, 44%, and 54%, respectively.

Estrogen induced hypertriglyceridemia and elevation ofplasma HDL cholesterol, phospholipid, and apolipoprotein A-Iand A-II concentrations. Low density lipoprotein (LDL) choles-terol decreased but LDL apolipoprotein B did not. These resultssuggest that the cardiovascular complications that occur duringestrogen administration are not mediated through changes inlipoprotein profile, other than the hypertriglyceridemic effect.Orchiectomy caused hypercholesterolemia and an increase inboth total and LDL apolipoprotein B, all of which are strongdeterminants of cardiovascular disease. The high density lipo-protein (HDL) concentration was not affected despite a reduc-tion in plasma testosterone, perhaps due to a simultaneousdecrease in estradiol. Combination therapy had no effect on

plasma lipid and apolipoprotein B concentrations, but very lowdensity lipoprotein (VLDL) apolipoprotein B decreased, andLDL apolipoprotein B increased. The HDL cholesterol andapolipoprotein A-I concentrations increased but A-II and phos-pholipids did not. These results suggest enhanced lipoproteinlipase activity, consistent with the reciprocal changes in VLDLand LDL apolipoprotein B levels, apolipoprotein B enrichmentof LDL particles, and increase in HDL cholesterol. The higherapolipoprotein A-I to A-II ratio indicates an increase in HDL2subfraction due to inhibition of endothelial hepatic lipase, in-creased secretion of apolipoprotein A-I, or both. These effectsare attributed to estradiol, which decreased less than after or-chiectomy, and to additional adrenal androgen inhibition byflutamide.

We conclude that estradiol plays an important role in deter-mining plasma lipoprotein concentrations in men, and andro-gens exert an antagonist effect. The lipoprotein profile resultingfrom the combination treatment is more beneficial than thatresulting from orchiectomy or estrogen administration. (J ClinEndocrinol Metab 66: 314, 1988)

CANCER of the prostate is among the most commonmalignancies in men and, unfortunately, is often

diagnosed after it has metastasized. Since the pioneeringwork of Huggins and Hodges (1), the treatment of dis-seminated prostatic carcinoma is aimed at limiting an-drogen-stimulated proliferation of the tumor, most oftenby orchiectomy or high doses of estrogens. Orchiectomyor estrogen treatment retards the progression of cancerin only 60-80% of patients; moreover, both produce

Received March 11,1987.Address all correspondence and requests for reprints to: Lipid Re-

search Unit, Le Centre Hospitalier de l'Universite Laval, 2705 Boule-vard Laurier, Quebec City, Quebec, Canada GlV 4G2.

cardiovascular complications (2, 3). In particular, in-creased mortality from adverse cardiovascular effects inestrogen-treated men was found in early coronary heartdisease drug trials (4).

Other ways of reducing androgen effects on prostaticcells have been sought and tested clinically. An alterna-tive approach has been the use of synthetic antiandro-genic compounds, such as cyproterone acetate (5). Stillmore recently, long-acting analogs of GnRH (LHRH)have been used alone (6, 7) to obtain castration plasmatestosterone levels or in combination with an antiandro-gen (8-10), which blocks the action of any remainingtestosterone. This latter approach has given promisingresults, improving the survival rate of patients with

314

ANDROGEN SUPPRESSION AND LIPOPROTEINS IN MEN 315

metastatic prostatic carcinoma from 42% to 65% after 3yr of treatment (10, 11).

Manipulations of sex hormone levels alter lipoproteinmetabolism (12). Because of the strong association be-tween lipoprotein levels and coronary heart disease it isimportant to investigate the effects on lipoproteins ofvarious treatments that alter sex hormone levels. Thisstudy focused on changes in plasma lipoprotein concen-trations during combination treatment with a LHRHagonist and an antiandrogen flutamide. Since data onthe comparative effects of various treatments on plasmalipoproteins and particularly apolipoprotein concentra-tions are lacking, patients with prior estrogen therapy ororchiectomy also were studied retrospectively.

Materials and MethodsPatient selection and study protocol

The study group consisted of 92 men, aged 49-86 yr. Theywere recruited consecutively from among men referred fortreatment of prostate cancer. Ninety were white caucassians(47 French Canadians and 43 Americans), and 2 were blackAmericans. The men were seen at the Laval University MedicalCenter out-patient clinic, informed written consent was ob-tained as required by the ethics committee, and history andphysical examination were performed as previously described(8, 9). All men had histologically proven adenocarcinoma of theprostate. Twenty-four had tumor localized to the prostate(stage C), while the remaining 68 had boney metastases (stageD2).

The study was conducted in two parts. The first was acomparison of plasma lipid and lipoprotein profiles in previ-ously untreated men with prostate cancer and men who hadundergone orchiectomy or were receiving diethylstilbestrol (3mg/day) at the time of referral (see Table 1). In the secondstudy, 27 previously untreated men received combinationtherapy consisting of LHRH agonist (D-Trp6,des-Gly-NH2

10)LHRH ethylamide (500 ng, sc, daily for 1 month, fol-lowed by a 250-Mg daily dose) and antiandrogen (flutamide or

TABLE 1. Characteristics of the patients according to treatment groups

its analog RU 23908; 250 mg, orally, three times a day) (8, 9).The patients taking combination therapy were resampled after6 months of treatment. They were not asked to change theirdiet. Blood samples were drawn into tubes containing EDTA(for lipid and lipoprotein measurements) and heparin (for ste-roid analyses) while the men were in a reclining position andafter a 12-h fast. Two blood samples were obtained on consec-utive days at the initial visit, whereas only a single sample wasobtained after the 6 months of treatment in the combinationtherapy group.

The mean (±SD) body mass index (BMI) of the group as awhole was 25.6 ± 3.6 kg/m2, which is lower than the reported50th percentile value of 27 in men 60-69 yr of age (13). Basedon the 10th and 90th percentiles, only 6 men were grosslyoverweight, whereas 9 were underweight. A weight loss or gainof more than 2 kg during the last 3 months before the initialvisit was reported by 13 and 7 men, respectively. The weightloss was voluntary in 4 men, being due to a low fat diet in 2men after detection of hypercholesterolemia and a macrobioticdiet in 2 other men. None of the patients had a history ofexcessive alcohol use. A majority of patients had maintainednormal activity, but 9 were symptomatic but ambulatory, in-cluding 2 who were bedridden for less than 50% of the time.None of the patients had liver, renal, or thyroid dysfunction.Seven patients had blood glucose levels above 130 mg/dL (7.22mmol/L), and 2 had elevated uric acid levels. Five patients hadhad a myocardial infarction in the past, 4 had stable angina,and 10 had hypertension (blood pressure, >160/90 mm Hg).The most frequently used medications were antihypertensivedrugs alone or in combination (/3-adrenergic antagonist drugsin 9 and thiazide diuretics in 12), oral hypoglycemic agents in4, and nitroglycerin in 5. The patients taking antihypertensivedrugs were approximately equally distributed among thegroups, and the medications were not changed in the combi-nation group studied prospectively.

Laboratory methods

Plasma testosterone and 170-estradiol levels were measuredin duplicate in plasma by double antibody RIAs, as describedpreviously (8). Plasma lipoprotein fractions were prepared by

Patients Untreated Estrogen Orchiectomy

Combination of LHRH agonistand flutamide

No.Age (yr)Wt (kg)Ht(m)BMI"Months of or after treat-

ment, mean (range)

44°66.7 ± 1.475.0 ± 2.11.72 ± 0.0125.2 ± 0.5

2565.1 ± 1.776.7 ± 2.81.76 ± 0.0224.8 ± 0.7

17.8(3.9-48)

2369.2 ± 1.382.8 ± 2.5*1.72 ± 0.0128.1±0.076'c'e

12.2(3.3-42)

Pre-Rx

2766.9 ± 1.874.0 ± 2.61.71 ± 0.01c

25.1 ± 0.7

Post-Rx

2767.4 ± 1.876.4 ± 2.71.71 ± 0.01c

26.1 ± 0.76

Values are the mean ± SEM.0 Including 27 patients in the combination treatment group.* Significantly different from the untreated group.c Significantly different from the estrogen-treated group.d Weight (kg)/height (m2).c Significantly different from the combination treatment (both pre- and post-) group.

316 MOORJANI ET AL. JCE & M • 1988Vol 66 • No 2

the combined use of ultracentrifugation (14) and heparin-man-ganese precipitation (15), as reported previously (16). Thesupernatant obtained after the heparin-manganese precipita-tion step containing high density lipoprotein (HDL) was usedto isolate HDL subclass 3, after precipitation of HDL subclass2 with dextran sulfate (mean mol wt, 15,000; SOCHIBO, Boul-ogne, France). The method of Gidez et al. (17) was used for theHDL subclass fractionation, but the final dextran sulfate con-centration was 1.9 g/L. The HDL2 cholesterol level was ob-tained by subtracting HDL3 cholesterol from HDL cholesterol.The cholesterol and triglyceride concentrations in plasma andlipoprotein fractions were quantified as previously described(18). The total phospholipid concentration (16) measured asphospholipid phosphorous was determined in plasma and theHDL fraction. The electroimmunoassay of Laurell (19) wasused to quantitate apolipoproteins. The total apolipoprotein Bconcentration was measured in plasma, whereas the densityfraction d > 1.006 g/mL was used, to measure low densitylipoprotein (LDL) apolipoprotein B levels. The very low densitylipoprotein (VLDL) apolipoprotein B concentration was ob-tained by difference. The apolipoprotein A and A-I levels weremeasured in the density fraction d > 1.006 g/mL, and theapolipoprotein A-II concentration was obtained by difference.Monospecific rabbit anti-apolipoprotein B, A, and A-I seraobtained from Behringwerke AG. (Marburg, West Germany)were used. The specificity of each antiserum was tested byradial double immunodiffusion, and the quantitative electroim-munoassays were carried out under standardized conditions, asdescribed previously (16, 20). The results in the untreated,estrogen-treated, and orchiectomy groups are the mean of theresults for the 2 daily blood samples, both analyzed at the sametime.

Statistical analyses

Statistical tests were performed as described by Snedcor andCochran (21). One-way analysis of variance was used to eval-uate differences among the various groups. Covariance analysiswas performed when significant Pearson's correlations werefound between confounding variables in pooled data from allpatients, both treated and untreated. The paired t test was usedto test the significance of differences between the pre- andposttreatment values. P < 0.05 was taken to indicate signifi-cance unless stated otherwise. The strengths of the relation-ships of sex steroids to other variables in the untreated groupwere estimated by the Spearman's rank correlation coefficient.

Results

The characteristics of the patients and the duration ofvarious treatments for all groups are shown in Table 1.The patients are not classified according to localized ormetastasized (boney metastases) disease status, sinceprevious results indicated that the plasma lipids, lipopro-teins, and apolipoproteins in patients stratified accordingto extent of disease were not different (16). The varioustreatment groups were similar, but the patients in theorchiectomy group were both heavier and had a higher

BMI. When the data from all groups, both treated anduntreated, were analyzed for the relationship of BMI(Pearson's correlation coefficient) to other variables, theresults indicated that BMI was weakly but significantlycorrelated to the following variables: plasma testoster-one, r = -0.23, P < 0.001; log 10 plasma triglycerides, r= 0.12, P < 0.001; phospholipids, r = 0.11, P < 0.05;apolipoprotein B, r = 0.12, P < 0.01; apolipoprotein A-II, r = -0.15, P < 0.001; and HDL3 cholesterol, r = -0.32,P < 0.001. These relationships were, therefore, takeninto consideration, and covariance analyses were per-formed to investigate the significance of differences inthese variables among the different treatment groups,with BMI as a covariate (see below).

Figure 1 shows the mean plasma testosterone and 17/3-estradiol concentrations in untreated men with prostaticcarcinoma and in those receiving treatment. The meanplasma testosterone level was markedly reduced in theorchiectomy group (by 91%) and in the medical castra-tion groups receiving either estrogen (by 89%) or com-bination therapy with LHRH agonist and flutamide (by92%). Similarly, mean plasma estradiol levels were alsoconsiderably lower, but the magnitude of the decreasewas variable; thus, the estradiol concentration was lowerby 85% in the orchiectomy group compared to 44% and54% in the estrogen and combination therapy groups,respectively.

The plasma lipid, lipoprotein fraction, and apolipopro-tein concentrations in the 27 patients treated with com-bination therapy are shown in Tables 2 and 3. Theplasma lipid and apolipoprotein B concentrations did notvary by more than 6% from the pretreatment value after6 months of treatment. The cholesterol concentration in

140

120

100-

80-

60

40

20

I

11

12-

10-

ocLUco 6O

co „LU 4

1

LI UNTREATEDtil ESTROGENS ORCHIECTOMYID LHRH-AGONIST

ANTIANDROGEN

n(100) (56) (15) (46) (100) (11) (9) (8)

FIG. 1. Mean (±SEM) plasma steroid concentrations during variousandrogen suppression treatments in men with prostatic carcinoma. Thevalues in parentheses are a percentage of the mean value in theuntreated group. The symbols represent the significance of the differ-ences from the untreated group (*), the estrogen group (f), the orchiec-tomy group (§), and the group receiving combination treatment withLHRH agonist and flutamide ($).

ANDROGEN SUPPRESSION AND LIPOPROTEINS IN MEN 317

TABLE 2. Effects of combined LHRH agonist and flutamide adminis-tration for 6 months on plasma lipids and apolipoprotein B levels in27 men with prostatic carcinoma

*

Cholesterol (mmol/L)

Triglycerides(mmol/L)

Phospholipids(mmol/L)

Apolipoprotein B(g/L)

Pretreatment

5.55 ± 0.18

1.62 ± 0.12

2.72 ± 0.09

1.13 ± 0.05

Treatment

5.69 ± 0.23(103)

1.56 ± 0.20(96)

2.87 ± 0.12(106)

1.19 ± 0.08(105)

P

NS

NS

NS

NS

Values are the mean ± SEM. The percent change from pretreatmentlevels is shown in parentheses.

TABLE 3. Effects of combined LHRH agonist and flutamide adminis-tration for 6 months on plasma lipoprotein lipids and apolipoproteinlevels in 27 men with prostatic carcinoma

VLDL°Cholesterol (mmol/L)

Triglycerides (mmol/L)

Apolipoprotein B (g/L)

LDLCholesterol (mmol/L)

Apolipoprotein B (g/L)

HDLCholesterol (mmol/L)

Phospholipids (mmol/L)

Apolipoprotein A-I(g/L)

Apolipoprotein A-II(g/L)

Pretreatment

0.46 ± 0.05

0.87 ± 0.09

0.17 ± 0.02

3.97 ± 0.15

0.97 ± 0.04

1.14 ± 0.05

1.11 ± 0.05

1.15 ± 0.04

0.45 ± 0.02

Treatment

0.41 ± 0.05(89)

0.68 ± 0.14(78)

0.12 ± 0.02(71)

3.90 ± 0.21(98)

1.06 ± 0.07(109)

1.37 ± 0.08(123)

1.16 ± 0.08(105)

1.42 ± 0.07(124)

0.48 ± 0.03(107)

P

NS

<0.05

<0.05

NS

<0.05

<0.05

NS

<0.05

NS

Values are the mean ± SEM. The percent change from pretreatmentlevels is shown in parentheses.

0 Log 10 values were used for statistical analyses of the VLDL lipidand apolipoprotein components.

both VLDL and LDL fractions did not change, in con-trast to a marked increase in the HDL fraction (Table3). The increase in the HDL cholesterol concentrationwas accompanied by a similar increase in apolipoproteinA-I, but not apolipoprotein A-II. The lack of significantincrease in plasma apolipoprotein B was due to thereciprocal changes in apolipoprotein B concentrations inthe VLDL (decreased) and LDL (increased) fractions.The VLDL triglyceride concentration decreased simi-larly.

Despite similarly low plasma testosterone concentra-tions in the three treatment groups, their plasma lipidand lipoprotein levels were variable (Tables 4 and 5).Compared to untreated patients, the estrogen-treated

patients had significantly higher plasma triglyceride andphospholipid levels (Table 4). The VLDL componentswere slightly but not significantly higher in the estrogen-treated men, as one would expect from an increase inplasma triglycerides (Table 5). The LDL cholesterolconcentration was significantly lower but the LDL apo-lipoprotein B concentration was not. The levels of allconstituents in the HDL fraction were higher; the rela-tive increase in apolipoprotein A-II was greater than thatin A-I. The significance of these results was not alteredwhen the data were adjusted for BMI as covariate, forlog 10 triglycerides, phospholipids, and apolipoproteinsA-II and B, all of which correlated significantly in variousdegrees to BMI.

The plasma cholesterol and apolipoprotein B concen-trations were higher in the orchiectomized patients thanin the estrogen-treated and untreated patients (Table 4),but the plasma triglyceride and phospholipid levels weresimilar. The LDL cholesterol and LDL apolipoprotein Bconcentrations also were higher in the orchiectomizedthan in the estrogen-treated patients (Table 5), but onlyin the latter were values significantly higher than thosein the untreated group, even after adjustment for BMI.However, the differences in the plasma apolipoprotein Blevels between the untreated and orchiectomy groupswere not significant. The concentrations of lipids andapolipoproteins in the VLDL and HDL fractions did notdiffer from those in the untreated patients.

When the results in the patients treated with LHRHagonist and flutamide for 6 months were compared tothose in all untreated patients by analysis of variance,the results (Tables 4 and 5) were similar to those shownfor the prospective study in Tables 2 and 3 and summa-rized above. The plasma cholesterol, triglyceride, andapolipoprotein B levels were similar to those in theuntreated group. However, plasma phospholipid levelswere significantly higher in the combination treatmentgroup (Table 4). As noted for orchiectomy, the differ-ences in LDL apolipoprotein B between the untreatedgroup and the combination treatment group were statis-tically significant in the covariate analysis with BMI ascovariate. In the VLDL fraction, the apolipoprotein Bconcentration was significantly lower, as also found inthe prospective study, but in contrast the VLDL choles-terol level, but not the VLDL triglyceride level, wassignificantly low.

The data on the relationships among sex steroids, age,BMI, and other measured lipoprotein variables in theuntreated group (n = 44) are shown in Table 6. The onlysignificant relationships for testosterone were its nega-tive correlation to BMI and positive correlation to HDLphospholipids. On the other hand, estradiol concentra-tion was strongly correlated to HDL2 cholesterol, but notto total HDL cholesterol or HDL3 cholesterol. None of

V

318 MOORJANI ET AL. JCE & M • 1988Vol66«No2

TABLE 4. Comparative effects of various androgen suppression treatments on the plasma lipid and apolipoprotein B concentrations in men withprostatic carcinoma

Cholesterol (mmol/L)

Triglycerides (mmol/L)c

Phospholipids (mmol/L)

Apolipoprotein B (g/L)

Untreated

5.43 ± 0.13

1.59 ± 0.09

2.56 ± 0.06

1.09 ± 0.04

Estrogen

5.22 ± 0.23(96)

2.16 ± 0.34(135)°

2.85 ± 0.10(112)°

1.02 ± 0.07(94)

Treated

Orchiectomy

5.97 ± 0.23(110)""

1.81 ± 0.12(113)

2.69 ± 0.09(105)

1.24 ± 0.06(114)°'*

LHRH agonist andflutamide

5.69 ± 0.23(105)

1.56 ± 0.20(98)"

2.87 ± 0.12(112)°

1.19 ± 0.08(109)*

Values are the mean ± SEM. The percentage of the mean value in the untreated group is shown in parentheses.0 Significantly different from the untreated group.6 Significantly different from the estrogen-treated group.c Log 10 values were used for the statistical analyses.

TABLE 5. Comparative effects of various androgen suppression treatments on the plasma lipoprotein-lipid and apolipoprotein concentrations inmen with prostatic carcinoma

Treated

UntreatedEstrogen Orchiectomy

LHRH agonist andflutamide

VLDLa

Cholesterol (mmol/L)Triglycerides (mmol/L)Apolipoprotein B (g/L)

LDLCholesterol (mmol/L)Apolipoprotein B (g/L)

HDLCholesterol (mmol/L)Phospholipids (mmol/L)Apolipoprotein A-I (g/L)Apolipoprotein A-II (g/L)

0.52 ± 0.050.79 ± 0.080.16 ± 0.01.

3.85 ± 0.130.93 ± 0.03

1.14 ± 0.051.08 ± 0.041.14 ± 0.040.48 ± 0.02

0.72 ± 0.18 (140)1.20 ± 0.24 (151)0.19 ± 0.03 (119)

3.18 ± 0.21 (83)M'C

0.86 ± 0.06 (92)"-e

1.34 ± 0.08 (118)M

1.33 ± 0.09 (123)M

1.46 ± 0.09 (128)6d

0.66 ± 0.04 (138)fcd<

0.59 ± 0.05 (115)0.91 ± 0.09 (116)0.16 ± 0.01 (100)

4.27 ± 0.21 ( l l l ) c

1.09 ±0.05 (117)*'

1.16 ± 0.05 (105)cc

1.08 ± 0.05 (100)c

1.14 ± 0.07 (100)c-e

0.46 ± 0.05 (96)c

0.41 ±0.05 (80)*•«•"0.68 ± 0.14 (86)c

0.12 ± 0.02 (75)6'c'd

3.90 ± 0.21 (101)c

1.06 ± 0.07 (114)6-c

1.37 ± 0.08 (121)6-rf

1.16 ± 0.08 (107)1.42 ± 0.07 (125)M

0.48 ± 0.03 (100)c

Values are the mean ± SEM. The percentage of the mean value in the untreated group is shown in parentheses." Log 10 values were used for the statistical analyses of the VLDL lipid and apolipoprotein components.6 Significantly different from the untreated group.c Significantly different from the estrogen-treated group.d Significantly different from the orchiectomy group.e Significantly different from the LHRH agonist and flutamide group.

the sex steroids correlated with HDL apolipoprotein A-Ior A-II levels. The ratio of testosterone to estradiolsignificantly correlated with both HDL2 cholesterol andHDL phospholipids, but these relationships were muchweaker than those for each steroid separately.

Discussion

Since the demonstration of the usefulness of LHRHagonists for medical castration therapy for the treatmentof prostatic carcinoma (7-9, 22), attention has beenfocused on the utility of LHRH agonists combined withantiandrogens (9-11) to prevent disease flares and im-prove patient survival by complete neutralization of an-drogens of both testicular and adrenal origin. However,

if patient survival is to be improved, antihormonal treat-ments intended for long term use not only have toprolong survival, but also have few metabolic side-effects.Sex hormones alter lipoprotein metabolism. Because ofthe strong association between lipoprotein levels andischemic heart disease, it is important to evaluate theeffects of antihormonal treatments on plasma lipopro-teins.

In the V.A. study (23), patients with prostate cancerhad lower serum cholesterol but similar triglyceride con-centrations compared to those in normal men. Althoughsimilar results were reported recently in a smaller groupof patients (24), the control population did not have thesame age distribution. In our patients, the cholesterolconcentrations in plasma and lipoprotein fractions were

y

ANDROGEN SUPPRESSION AND LIPOPROTEINS IN MEN 319

TABLE 6. Spearman's correlations in 44 untreated men with prostaticcarcinoma: sex hormones, lipoproteins, and apolipoproteins

Age (yr)BMICholesterolTriglyceridesPhospholipidsApolipoprotein BApolipoprotein A-IApolipoprotein A-IIVLDL cholesterolVLDL apolipoprotein BLDL cholesterolLDL apolipoprotein BHDL cholesterolHDL2 cholesterolHDL3 cholesterolHDL phospholipids

Estradiol

0.21-0.06-0.07-0.02

0.03-0.14

0.23-0.03

0.200.07

-0.09-0.17

0.080.52°

-0.010.15

Testosterone

0.12-0.31*

0.090.080.250.040.110.240.180.140.09

-0.020.110.060.230.36*

Testosterone/estradiol ratio

-0.05-0.12

0.150.150.230.220.080.210.110.030.140.160.200.38*0.250.27*

°P<0.01.*P<0.05.

similar to the 50th percentile of the age group 65-69 yrreported by Lipid Research Clinics Program (25). Also,the results were similar to our previously reported find-ings in prostate cancer as well as normal subjects (16).In the latter study we found that plasma lipoproteinlevels were similar in patients with stage C or metastaticD2 stage of the disease.

Estrogen administration significantly increasedplasma triglyceride levels, as reported previously (24, 26,27). These patients had been receiving estrogen for anaverage of 18 months, and therefore, it may be arguedthat the hypertriglyceridemic response in these patientsmay not be representative of all estrogen-treated pa-tients, since excess deaths usually occur during the initialperiod of estrogen therapy. However, the 35% increasein plasma triglyceride levels is similar to the increasesreported in men treated for 1-2 months (24, 26, 27), butconsiderably less than the 60% increase in men treatedfor 6 months (24). The mean plasma cholesterol concen-tration was similar in the estrogen-treated and untreatedmen, as reported previously (24, 26, 27), as was plasmaapolipoprotein B. However, LDL cholesterol was signif-icantly lower and HDL cholesterol significantly higher.The increase in HDL cholesterol concentration was ac-companied by parallel increases in HDL phospholipidand apolipoproteins A-I and A-II, thus indicating anincrease in HDL mass and/or particle number. Thesechanges are similar to those reported previously inwomen (28). The alterations in plasma lipoproteins dur-ing estrogen therapy cannot be attributed to suppressionof endogenous testosterone secretion as shown by theresults in the orchiectomy group and probably are due toestrogen-induced changes in hepatic metabolism (29)

leading to simultaneous increases in hepatic VLDL andHDL secretion (30). Decreased hepatic endothelial lipaseactivity may also contribute to an increase in HDLcholesterol concentration, as reported previously forwomen taking estrogenic oral contraceptives (31). Theinflux of triglycerides is most likely enhanced (32). Stim-ulation of hepatic LDL receptors (33) may partiallyexplain the decrease in LDL; moreover, the lesserchanges in LDL and VLDL apolipoprotein B comparedto cholesterol levels suggest that enhanced triglyceridesecretion may be accompanied by less markedly in-creased influx of apolipoprotein B compared to otherapolipoproteins, as reported in cultured cells (34).

Orchiectomy did not alter plasma triglyceride orVLDL, LDL, and HDL cholesterol levels, but plasmacholesterol was significantly higher than that in un-treated men, as documented previously (24, 26). Plasmaapolipoprotein A-I and A-II and HDL phospholipid levelswere similar in the two groups, indicating that HDLmetabolism was not affected despite the much lowertestosterone levels in the orchiectomy group. On theother hand, consistent with the higher plasma cholesterollevel, plasma and LDL apolipoprotein B concentrationswere significantly higher. These results were somewhatunexpected and do not readily explain why men havehigher LDL cholesterol concentrations than women. Asimilar phenomenon has been reported during LHRHagonist administration in normal young men (35). Themean BMI of the orchiectomy group was significantlyhigher than that of the untreated group as well as of theother treatment groups, and since BMI was weakly butsignificantly correlated to apolipoprotein B, the covariateanalysis showed (P < 0.12) that the plasma apolipopro-tein B increase is partially explained by excess BMI.Since orchiectomy reduced plasma estradiol as well asplasma testosterone, the estradiol decrease may havecaused the elevation of LDL apolipoprotein B and thelack of change in HDL fraction. Taken together theseresults indicate that after orchiectomy, androgens ofadrenal origin have some influence on LDL and HDLmetabolism. Furthermore, the BMI increase may be dueto adrenal androgens, based on previous observations inprepubertal boys (36), in whom elevated adrenal andro-gen levels are associated with increased body fat.

The combined administration of LHRH agonist andflutamide, in contrast to orchiectomy, resulted in plasmacholesterol levels similar to those in untreated men. Also,in contrast to orchiectomy, the plasma apolipoprotein Bconcentration did not increase significantly, althoughthe LDL and VLDL apolipoprotein B levels were, re-spectively, higher and lower. These results suggest thatperipheral lipoprotein lipase may be more active duringcomplete androgen neutralization achieved by suppres-sion of testicular androgen secretion by LHRH agonist

320 MOORJANI ET AL. JCE & M • 1988Vol 66 • No 2

and competitive inhibition of androgen action by fluta-mide. Although the plasma triglyceride concentrationwas unaltered, despite what would be expected fromincreased lipoprotein lipase activity, smaller decreases inboth plasma triglycerides and VLDL triglycerides (16,37) have been found early during such treatment. More-over, the reduction in VLDL apolipoprotein B concen-tration also indicates stimulation of VLDL hydrolysisand further explains the accumulation of LDL particlesbecause of a precursor-product relationship and probablyfrom a lack of increase in LDL catabolism. The moststriking effect of combination treatment was the higherHDL cholesterol and apolipoprotein A-I concentrations,in contrast to the lack of similar HDL changes afterorchiectomy (24, 26) or during LHRH agonist therapyalone (37). The increases in HDL concentrations duringcombination treatment may have resulted from a directaction of flutamide on HDL metabolism or indirectlythrough inhibition of adrenal androgen action or due tolesser estradiol reduction. The similar decrease in plasmatestosterone after orchiectomy and its lack of effect onall HDL components, as found by us and also reportedpreviously for HDL cholesterol (24, 26, 37), compared tocombination treatment, suggest that the higher HDLcholesterol and apolipoprotein A-I levels during the latterwere due to the lesser estradiol decrease, absence ofadrenal androgen action, or both. The possibility ofincreased lipoprotein lipase activity, based on the recip-rocal changes in the VLDL and LDL apolipoprotein Blevels, suggests that the increase in VLDL hydrolysismay have contributed to the higher HDL cholesterol andparticularly HDL2 cholesterol levels. However, such anincrease may also result from a decrease in HDL2 catab-olism by inhibition of hepatic lipase (31), as well asincreased production of apolipoprotein A-I (30). Both ofthese effects are most likely due to endogenous estradiolin the absence of antagonism from androgens, as dis-cussed above.

It is important to point out that all antiandrogens donot have similar effects on plasma lipoprotein metabo-lism. Thus, contrary to the effects of flutamide, cypro-terone acetate reduces both plasma and HDL cholesterolconcentrations (37-39). Such reductions are not due tothe antiandrogen properties of cyproterone acetate, butare more likely the result of other androgenic or proges-tagenic properties of cyproterone acetate (40). Otherandrogenic steroids and compounds with progestagenicactivity have similar effects in men (41).

Sex differences in plasma lipoprotein concentrationsare usually explained on the basis of ovarian function(12). However, longitudinal data from adolescent menindicate that HDL cholesterol concentrations decrease

during sexual maturation (42), thereby suggesting theimportance of androgens in determining HDL levels inadult men. From the foregoing discussion it is apparentthat endogenous estradiol elevates HDL concentrationsin men, although the extent of its effect is probablydependent on the antagonist action of androgens. Thisview is also supported by our data on the strong associ-ation between estradiol and HDL2 cholesterol concentra-tions in untreated men with prostatic carcinoma. Fur-thermore, this relationship is weaker when the testoster-one concentration is used as a denominator of estradiol.Paradoxically, cross-sectional studies of adult men havegenerally found a positive association between plasmatestosterone and HDL cholesterol levels (43-46); how-ever, these relationships were weak. A negative associa-tion between plasma free testosterone concentration andHDL cholesterol was reported in one study (47).

The main clinical relevance of changes in circulatinglipids and apolipoproteins is their relation to cardiovas-cular risk. Except for hypertriglyceridemia during estro-gen therapy, other changes in the lipoprotein profile canbe interpreted as favorable. Nevertheless, the clinicalbenefits of simultaneous increases in plasma HDL andtriglycerides are not clear. The high mortality from car-diovascular diseases that occurs initially as well as duringlong term treatment with estrogen (3,22) does not appearto be due to changes in lipoprotein profile per se. Orchiec-tomy alone results in small but significant increases inplasma cholesterol as well as plasma and LDL apolipo-protein B concentrations, all of which are importantdeterminants of coronary heart disease. The patientstreated with combination therapy with LHRH agonistand antiandrogen flutamide (without androgenic prop-erties) had significantly higher HDL concentrations. Al-though the LDL apolipoprotein B concentration also washigher, the relative increase was small. Thus, the com-bination antihormonal therapy may be safer for longterm management of prostatic cancer patients.

Acknowledgments

We are grateful to the staff of the Lipid Research Unit for theirtechnical assistance, to Mrs. Suzanne Croft, R.N., for obtaining bloodsamples and project coordination, and Mrs. Danielle Lemay and HeleneBoucher for their valuable assistance in preparation of the manuscriptand statistical analyses, respectively.

References1. Huggins C, Hodges CV 1941 Studies of prostatic cancer. I. Effect 4

of castration, estrogen and androgen injections on serum phospha-tases in metastatic carcinoma of the prostate. Cancer Res 1:293

2. Veterans Administration Cooperative Urological Research Group1967 Carcinoma of the prostate: treatment comparisons. J Urol98:516

ANDROGEN SUPPRESSION AND LIPOPROTEINS IN MEN 321

3. Blackard CF, Doe RP, Mellinger GT, Byar DP 1970 Incidence ofcardiovascular disease and death in patients receiving diethylstil-boesterol for carcinoma of the prostate. Cancer 26:249

4. Coronary Drug Project Research Group 1973 The coronary drugproject. Findings leading to discontinuation of the 2.5 mg estrogengroup. JAMA 226:652

5. Wein AJ, Murphy JJ 1973 Experience in treatment of prostaticcarcinoma with cyproterone acetate. J Urol 109:68

6. Waxman JH, Wass JAH, Hendry WF, Whitfield HN, Besser GM,Malpas JS, Oliver RTD 1983 Treatment with gonadotropin releas-ing hormone analogue in advanced prostatic cancer. Br Med J286:1309

7. Tolis G, Ackman D, Stellos A, Mehta A, Labrie F, Fazekas ATA,Comaru-Schally AM, Schally AV 1982 Tumor growth inhibitionin patients with prostatic carcinoma treated with LHRH agonists.Proc Natl Acad Sci USA 79:1658

8. Labrie F, Dupont A, Belanger A, Causan L, Lacourciere Y, Mon-fette G, Laberge JG, Emond JP, Fazekas ATA, Raynaud JP,Husson JM 1982 New hormonal therapy in prostatic carcinoma:combined treatment with an LHRH-agonist and an antiandrogen.Clin Invest Med 5:267

9. Labrie F, Dupont A, Belanger A 1985 Complete androgen blockadefor the treatment of prostatic cancer. In de Vita Jr VT, HellmanS, Rosenberg AS (eds) Important Advances in Oncology. Lippin-cott, Philadelphia, p 193

10. Klijn JGM, de Voogt HJ, Schroder FH, de Jong FH 1985 Combinedtreatment with buserelin and cyproterone acetate in metastaticprostatic carcinoma. Lancet 2:493

11. Labrie F, Dupont A, Belanger A, Poyet P, Giguere M, LacourciereY, Emond J, Monfete G, Borsanyi JP 1986 Combined treatmentwith flutamide and surgical or medical (LHRH agonist) castrationin metastatic prostatic cancer. Lancet 1:49

12. McGill HC, Stern MP 1979 Sex and atherosclerosis. In Paoletti R,Gotto AM (eds) Atherosclerosis Reviews. Raven Press, New York,vol 5:157

13. Health and Welfare Canada/Statistics Canada 1981 The Health ofCanadians: Report of the Canada Health Survey. Minister ofSupply and Services, Ottawa, Canada, catalog no. 82-538

14. Havel RJ, Eder HA, Bragdon HA 1955 The distribution andchemical composition of ultracentrifugally separated lipoproteinsin human serum. J Clin Invest 34:1345

15. Burstein M, Samaille J 1960 Sur un dosage rapide du cholesterollie aux /9-lipoproteines du serum. Clin Chim Acta 5:609

16. Moorjani S, Dupont A, Labrie F, Lupien PJ, Brun D, Gagne C,Giguere M, Belanger A 1987 Increase in plasma high densitylipoprotein concentration following complete androgen blockage inmen with prostatic carcinoma. Metabolism 36:244

17. Gidez LI, Miller GJ, Burstein M, Slagle S, Eder HA 1982 Separa-tion and quantitation of subclasses of human plasma high densitylipoproteins by a simple precipitation procedure. J Lipid Res23:1206

18. Moorjani S, Gagne C, Lupien PJ, Brun D 1986 Plasma triglyceridesrelated decrease in high-density lipoprotein cholesterol and itsassociation with myocardial infarction in heterozygous familialhypercholesterolemia. Metabolism 35:311

19. Laurell CB 1972 Electroimmunoassay. Scand J Clin Lab Invest[Suppl 124] 29:21

20. Curry MD, Alaupovic P, Suennram CA 1976 Determination ofapolipoprotein A, A-I and A-II polypeptides by separate electroim-munoassays. Clin Chem 22:315

21. Snedecor GW, Cochran WG 1980 Statistical Methods, ed 7. IowaState University Press, Ames, pp 135-174

22. Leuprolide Study Group 1984 Leuprolide versus diethylstilbestrolfor metastatic prostate cancer. N Engl J Med 311:1281

23. Seal US, Doe RP, Byar DP, Corle DK, V.A. Cooperative UrologicalResearch Group 1976 Response of serum cholesterol and triglyc-erides to hormone treatment and the relation of pretreatmentvalues to mortality in patients with prostatic cancer. Cancer38:1095

24. Rossner S, Hedlund PO, Jogestrand I, Sawe U 1985 Treatment ofprostatic cancer: effects on serum lipoproteins and the cardiovas-cular system. J Urol 133:53

25. U.S. DHHS 1980 The Lipid Research Clinics Population StudiesData Book, vol 1. NIH Bethesda, publication no. 80-1257

26. Wallentin L, Varenhorst E 1981 Plasma lipoproteins during anti-androgen treatment by estrogens or orchidectomy in men withprostatic carcinoma. Horm Metab Res 13:293

27. Bulusu NV, Lewis SB, Das S, Clayton Jr WE, 1982 Serum lipidchanges after estrogen therapy in prostatic carcinoma. Urology20:147

28. Cheung MC, Albers JJ 1977 The measurement of apolipoproteinAI and All levels in men and women by radioimmunoassay. J ClinInvest 60:43

29. Ottosson UB, Carlstrom K, Johansson BG, Von Schaultz B 1986Estrogen induction of liver proteins and high-density lipoproteincholesterol: comparison between estradiol valerate and ethinylestradiol. Gynecol Obstet Invest 22:198

30. Schaefer EJ, Foster DM, Zech LA, Lindgren FT, Brener MB, LevyRI1983 The effects of estrogen administration on plasma lipopro-tein metabolism in premenopausal women. J Clin Endocrinol Me-tab 57:262

31. Tikkanen MJ, Nikilla EA, Kuusi T, Sipinen S 1982 High densitylipoprotein-2 and hepatic lipase: reciprocal changes produced byestrogen and norgestrel. J Clin Endocrinol Metab 54:1113

32. Glueck C, Fallat RW, Scheel D 1975 Effects of estrogenic com-pounds on triglyceride kinetics. Metabolism 24:537

33. Chao Y, Windier EE, Chen GC, Havel RJ 1979 Hepatic catabolismof rat and human lipoproteins in rats treated with 17-ethinylestradiol. J Biol Chem 254:11360

34. Tam SP, Archer TK, Deeley RG 1985 Effects of estrogen onapolipoprotein secretion by the human hepatocarcinoma cell line,Hep G2. J Biol Chem 260:1670

35. Goldberg RB, Rabin D, Alexander AN, Doelle GC, Getz GS 1985Suppression of plasma testosterone leads to an increase in serumtotal cholesterol and high density lipoprotein cholesterol and apo-lipoproteins A I and B. J Clin Endocrinol Metab 60:203

36. Katz SH, Hediger ML, Zemel BS, Parks JS 1985 Adrenal andro-gens, body fat, and advanced skeletal age in puberty: new evidencefor the relations of adrenarche and gonadarche in males. Hum Biol57:401

37. Paisey RB, Kadow C, Bolton C, Hartog M, Gingell JC 1986 Effectsof cyproterone acetate and a long-acting LHRH analogue on serumlipoproteins in patients with carcinoma of the prostate. J R SocMed 79:210

38. Wallentin L, Varenhorst E 1980 Plasma lipoproteins during treat-ment with cyproterone acetate in men with prostatic carcinoma. JClin Endocrinol Metab 51:1118

39. Wynn V, Godsland IF, Seed M, Jacobs HS 1986 Paradoxical effectsof the antiandrogen cyproterone acetate on lipid and lipoproteinmetabolism. Clin Endocrinol (Oxf) 24:183

40. Poyet P, Labrie F 1985 Comparison of the antiandrogenic/andro-genic activities of flutamide, cyproterone acetate and megestrolacetate. Mol Cel Endocrinol 42:283

41. Webb OL, Laskarzewski PM, Gleuck CG 1984 Severe depressionof high density lipoprotein cholesterol levels in weight lifters andbody builders by self-administered exogenous testosterone andanabolic-androgenic steroids. Metabolism 33:971

42. Morrison JA, Laskarzewski PM, Rauh JL, Brookman R, MelliesM, Frazer M, Khoury P, de Groot I, Kelly K, Gleuck CJ 1979Lipids, lipoproteins and sexual maturation during adolescence: thePrinceton Maturation Study. Metabolism 28:641

43. Nordoy A, Aakvagg A, Thelle D 1979 Sex hormones and highdensity lipoproteins. Atherosclerosis 34:431

44. Gutai J, Laporte R, Kuller L, Dai W, Falvo-Gerard L, Caggiula A1981 Plasma testosterone, high density lipoprotein cholesterol andother lipoprotein fractions. Am J Cardiol 48:897

45. Heller RF, Wheeler MJ, Micallef J, Miller NE, Lewis B 1983

322 MOORJANI ET AL. JCE&M-I988Vol66»No2

Relationship of high density lipoprotein cholesterol with total and and to the binding capacity of sex hormone binding globulin infree testosterone and sex hormone binding globulin. Acta Endocri- healthy Finnish men. Metabolism 35:535nol (Copenh) 104:253 47. Semmens J, Rouse I, Beilin LJ, Masarei JRL 1983 Relationship of

46. Hamalainen E, Adlercreutz H, Ehnholm C, Puska P 1986 Rela- plasma HDL cholesterol to testosterone, estradiol, and sex-hor-tionships of serum lipoproteins and apoproteins to sex hormones mone-binding globulin in men and women. Metabolism 32:428

Postgraduate Course in Gynecologic and Obstetric Pathology

A postgraduate course in Gynecologic and Obstetric Pathology, sponsored by the Departments of Pathology,Massachusetts General Hospital, Brigham and Women's Hospital, and the Harvard Medical School, willbe held April 11 to 15, 1988 in Boston, MA. The course is designed for pathologists and obstetrician-gynecologists at the resident and practitioner levels, and will provide an in depth review with emphasis onmorphologic diagnostic features and clinicopathologic correlations. Special attention will be paid to recentadvances and newly recognized entities. The course is accredited by the AMA and the ACOG. The fee forthe course is $595 ($395 for residents and fellows).

For further information please contact:

Department of Continuing EducationHarvard Medical School25 Shattuck StreetBoston, MA 02115

Telephone:(617) 732-1525