Effect of Physical Training on Lipids, Lipoproteins, Apolipoproteins, Lipases, and Endogenous Sex Hormones in Men With Premature Myocardial Infarction Soaira G. Mendoza, H. Carrasco, A. Zerpa, Y. Briceno, F. Rodriguez, J. Speirs, and C.J. Glueck In 17 men, aged 27 to 54 years, with myocardial infarction 2 to IO months before the current exercise study, we aimed to determine whether 3 months of exercise training, at a level designed to elevate high-density lipoprotein cholesterol (HDLC), would be associated with changes in endogenous sex steroid hormones and postheparin lipoprotein and hepatic lipases, and whether the changes in sex hormones, lipids, lipoproteins, apolipoproteins, and physical activity were interrelated. Supervised bicycle ergometry, 30 minutes, 3 days per week, eliciting 75% of maximum heart rate, produced a significant training effect, with a 26% increase in the duration of the exercise test at a standardized, submaximal workload (P 5 .OOl), and a reduction in heart rate measured at a standardized submaximal workload, P = .06. After 3 months’ training, mean HDLC increased 23% (30 to 37 mg/dL), P I .OOl, mean apo A2 increased 19% (43 to 51 mg/dL), P I .OOl, and the ratio of total cholesterol (TC) to HDLC decreased 26% (P < .Ol), while estradiol (E,) levels decreased 45% (50.1 to 27.6 pg/mL). P I .OOOl. After 1 and 2 months’ exercise, TC (12% [P 2 .OOl]. 11% [P 5 .Ol]). and low-density lipoprotein cholesterol (LDLC) (13% [P I .Ol). 12% {P 5 .Ot)) were reduced. Hepatic lipase decreased 16% (P < .Ol) and 16% (P I .05) after 1 and 3 months’ exercise. There were no significant changes in apo Al, lipoprotein lipase, testosterone, luteinizing hormone (LH), follicle-stimulating hormone (FSH), or weight. By stepwise regression analysis, after 3 months' training, 66% (P = .0025) of the variance for the increase in HDLC from baseline to day 90 was accounted for independently by a decrease in triglyceride (F = 13.2, P = .003), by reduced heart rate on a fixed submaximal load (F = 12.7, P = .0035), and by a decrease in hepatic lipase (F = 5.5, P = .036). A modest, achievable exercise program can have significant cardiovascular benefit for men after myocardial infarction by ameliorating their hyperestrogenemia, reducing TC and LDLC, improving the TC to HDLC ratio, and elevating HDLC and apo A2. The increment in HDLC was related independently to improved capacity to sustain submaximal exercise and to exercise-induced reductions in triglyceride and postheparin hepatic lipase. Copyright o 1991 bbyW.B. Saunders Company

B

ASED ON extensive epidemiologic and controlled clinical trial data, the National Consensus Conference on lowering cholesterol and the National Cholesterol Education Program have recommended total cholesterol (TC) screening in all adults and lipoprotein cholesterol screening in subjects at elevated cardiovascular risk, as well as a concerted plan of intervention.‘” Moreover, the importance of high-density lipoprotein cholesterol (HDLC) as an independent, inverse risk factor for coronary heart disease has become better appreciated.3-7 Also, the Helsinki Heart Trial showed that diet and Lopid therapy, which increased HDLC and reduced low-density lipoprotein cholesterol (LDLC) and triglyceride levels, also significantly reduced coronary heart disease events.8.’ There is increasing emphasis on the importance of dealing with low HDLC and high TCHDLC as an independent lipoprotein risk factor.4-‘2 Factors that profoundly affect HDLC levels are heritability, age, the degree of sexual maturation, cigarette smoking, alcohol intake, ponderosity, sex, and race, as well as habitual and leisure time physical activity.‘“~‘6

From the Endocrinology, Cardiology, and Physical Rehabilitation units of University of the Andes, Merida, Venezuela, and the Cholesterol Center, Jewish Hospital, Cincinnati, OH. This research followed a protocol approved by the Institutional Review Committee, University ofAndes, with signed informed consent. Supported by Grant Conicit Sl-1555, University of the Andes, and by the Jewish Hospital Medical Research Council. Address reprint requests to Charles Glueck, MD, Cholesterol Center, Jewish Hospital, 3200 Bumet Ave, Cincinnati, OH 45229. Copyright Q 1991 by KB. Saunders Company 0026-0495/91/4004-0007$03.0010

The inverse association between physical activity and cardiovascular disease incidencc”~‘* appears to relate, at least in part, to the effects of physical activity on HDL~.13.14.19-23 Controlled trials of increased physical activity in normal men have usually shown increases in HDLC levels,‘9~2* but there have been variable changes in women’s HDLC levels in physical exercise studies.‘9~23~24 It is possible that male-female differences in HDLC response to exercise could be related to different changes in endogenous sex hormones during the exercise training periods,‘S.2h and to the fact that premenopausal women start with a higher pretraining HDLC than men. Physical training can change endogenous hormone levels; prolonged exercise has been reported to be associated with both a reduction in plasma testosterone levels,‘6 and an increase.” Endogenous sex hormones are potent regulators of plasma lipoprotein levels.27-33Endogenous plasma testosterone levels are positively related to plasma HDLC in menz7-32 and in oligospermic and azoospermic subjects.33 A high incidence of hyperestrogenemia has been reported in young men who had suffered myocardial infarction.‘7.LY.3” Recently, we reported” that estradiol (E,), the E, to testosterone ratio, and apo Al and B are important, independent markers of atherosclerotic coronary artery disease in young men with premature myocardial infarction. In the current study, our specific aim was to determine whether exercise training in men with premature myocardial infarction, at a level designed to elevate HDLC, would be associated with changes in endogenous sex steroid hormones, and postheparin lipoprotein and hepatic lipases, and whether the changes in sex hormones, lipids, lipoprotein cholesterols, apolipoproteins, and physical training outcome were interrelated.

Metabolism, Vol40, No 4 (April), 1991: pp 368-377

369

TRAINING, LIPOPROTEINS, SEX HORMONES

MATERIALS

AND METHODS

Subjects and Study Conditions for Subjects Seventeen male patients, aged 27 to 54 (42.2 + 7.9) years were studied. All subjects participated with signed informed consent. They had sustained myocardial infarction 2 to 10 months before the study. Because change in diet (reduced fat, low cholesterol) may affect serum E, levels,J4 and because male smokers have been reported to have high E2 levels,” and because drugs may affect E? levels,” we made every effort, where possible, to control these exogenous factors before starting the training program, so that they would not affect any decrements in E, observed during exercise. Twelve of the 17 patients (71%) smoked before their myocardial infarctions. These 12 smokers stopped cigarettes at least 2 to 3 months before initiating the exercise program. No subjects were smoking during the physical training period. During the 3-month period of physical training, the subjects were not given any special dietary restrictions in cholesterol or saturated fat. The patients were urged not to lose weight during the training program and there were no significant differences (P > .l) between baseline, pre-exercise program weight, and weight after 90 days (Table 1). No body composition or abdominal fat measurements were performed. No specific restrictions on alcohol intake were enforced, although subjects were encouraged to follow their habitual alcohol intake. Seven of the 17 patients (41%) were mild hypertensives (diastolic blood pressure > 100 mm Hg), controlled with a salt-restricted diet. These seven patients were maintained throughout the study on the same salt-restricted diets they had followed for 6 or more months before physical training. None of the patients changed their diets within 2 months of, or at the start of, or during the exercise period. Three of the 17 patients had started P-blocker adrenergic drugs for control of hypertension 2 or more months before beginning physical training, and their P-blocker therapy was unchanged throughout the exercise period. Changes in Ez levels in the three

subjects receiving P-blockers were no different during the exercise period (P > .l) than in the 14 subjects not receiving P-blockers. Thus, we believe that it is unlikely that exogenous factors, including diet, smoking, or drugs. affected either the decrement of E2 dcring physical training or other aspects of the study outcome. The patients had no other serious disorders besides previous myocardial infarction and diet-controlled mild hypertension (n = 7). Twelve of the 17 patients (71%) had angiocardiograms (Judkins procedure”) with the following results: one-vessel disease in six patients, two-vessel disease in four patients, three-vessel disease in two patients. Stress electrocardiography using a treadmill (Bruce protocol”) was performed both before and after the physical training program; none of the patients showed ischemic changes. Any drug that could interfere with the stress electrocardiography was discontinued 24 hours before the test was performed. When the patients reached maximal heart rate for their ages” or when they had any symptoms, the exercise treadmill test was suspended. VOZ was not measured; its measurement is not included in the University of the Andes, Merida cardiac rehabilitation program. We measured two parameters of cardiorespiratory fitness: heart rate measured at a standardized submaximal workload, and total time of exercise expended to exhaustion at a standardized submaximal workload. These measurements were made immediately before and after 90 days of physical training. We also classified cardiovascular capacity, using the categories of Astrand and Rylonint.” which are based on heart rate achieved at a standardized submaximal workload. Cardiorespiratory (aerobic) capacity represents the ability to sustain prolonged submaximal exercise, whereas maximal 0, consumption (maximal aerobic power) represents the “power” of the cardiorespiratory system.

Study Protocol After completing above), the subjects

their entry exercise electrocardiogram (as participated in a physical training program at

Table 1. Effects of Exercise on Mean (SD) Lipids, Lipoproteins, Apolipoproteins, Lipolytic Enzymes, and Endogenous Sex Hormones in Postmyocardial Infarction Men Baseline

Day 30

Pre-exercise

on Exercise

on

Day 60

Exercise

Day 90

on Exercise

Triglyceride (mg/dL)

311 (197)

231 t (95)

228t (70)

236* (78)

TC lmg/dL)

222 (34)

195* (30)

197$ (27)

214 (26) 37* (10)

HDLC (mg/dL)

30 (90)

33* (8)

37$ (9)

LDLC (mg/dL)

132 (32)

115t (34)

116t (31)

128 (25)

TC/HDLC

8.1 (3.0)

6.2t (1.6)

5.6* (1.4)

6.0t (1.5)

VLDL triglyceride (mg/dL)

146 (81)

114’144)

112t (35)

120* (44)

VLDL cholesterol (mg/dL)

48 (26) 96 (78)

42 (20)

40* (17)

44 (18)

73 (31)

71* (30)

67* (20)

Apo Al (mg/dL)

119 (22)

123 (24)

121 (18)

118 (23)

Apo A2 lmg/dL)

43 (9) 117 (19)

47 (9) 110 (15)

50* (8)

51* (7)

108* (12)

113 (15)

38 (13)

32t (11)

33 (13)

32* (12)

22 (7) 50.1 (15)

21.0 (10) 34.1* (14)

21 (7) 27.5* (11)

22 (8) 27.8t (11)

VLDL protein (mg/dL)

Apo B (mg/dL) Hepatic lipase (pm FFA/mL/h) Lipoprotein lipase (km FFAImUh) Estradiol (pg/mL) Testosterone (ng/mL)

7.8 (2)

7.7 (2)

7.2 (2)

7.4 (2)

LH (mlU/mL)

9.9 (4)

9.5 (4)

8.8 (3)

8.8 (3)

FSH (mlU/mL)

8.3 (5)

8.6 (4)

9.0 (4)

7.5 (4)

Prolactin (mg/dL)

6.6 (3)

8.0 (3)

8.4* (4)

8.4* (3)

68.8 (9)

68.2 (9)

68.4 (9)

Weight (kg) Maximal exercise heart ratelmin Exercise time (min)

68.9 (9)

169.4 (12)

163.7 (12)

10.8 (2)

13.6* (2)

All comparisons Y baseline, two-way ANOVA: *P 5 .05; tP 5 .Ol; SP I ,001,

370

the physical rehabilitation unit of the University of the Andes, Merida, Venezuela (altitude 5,400 feet). The patients participated in 3 months of supervised physical training, using a cycle ergometer, exercising for 30 minutes, 3 days per week, at exercise intensities eliciting 75% of maximum heart rate. The heart rate was continuously monitored during the cycle-ergometer workouts. Absences from sessions were made up, when possible, in the same week as the missed session or in the subsequent week. Attendance was good throughout, as good in the last 30 days as in the first 30 days. As summarized in Table 1, after a 12-hour fast, subjects were sampled for lipids, lipoproteins, apolipoproteins, postheparin hepatic and lipoprotein lipase, and endogenous sex steroid hormones. These studies were performed at baseline (pre-exercise), and then at days 30, 60. and 90 during exercise. Weight was measured at each visit. Analytical Techniques Fasting (12 to 14 hours) blood samples were drawn from an antecubital vein and collected in tubes containing 1 mg/mL of the disodium salt of EDTA. Blood samples were collected 24 hours after the last training session. Plasma samples were analyzed for total cholesterol and triglycerides by an enzymatic procedure using the ABA-50 autoanalyzer with an automated analysis reagent kit and methods supplied by Boehringer, Lewis, UK. HDLC levels were measured by the heparin-manganese precipitation method.39 LDLC was calculated by the Friedewald formula:” In samples with triglyceride greater than 300 mg/dL, a direct quantitation of LDLC was obtained by isolating the d > 1.006 fraction. The TC in this fraction was measured and heparin manganese added to another aliquot to precipitate LDL. HDLC was measured directly in the supernate. Very-low-density lipoproteins (VLDL) for measurement of VLDL triglyceride and protei@ were separated by ultracentrifugation in all subjects using a Beckman LS-50 centrifuge (Fullerton, CA) and 40.3 rotor following centrifugation at 40,000 rpm at 15°C for 18 hours. Apo B was quantitated by an immunoelectrophoresis assay following the procedure described by Cheung and Alber? and LaurelLJ Total plasma apo Al and A2 were measured by the immunoelectrophoresis assay similar to that for apo B using monospecific antiserum against these proteins, obtained by isoelectrofocusing and column chromatography. The interassay and intraassay coefficients of variation for both methods were less than 5%. Testosterone and ELwere measured by specific radioimmunoassay using commercial kits (Diagnostic Product Corp, Los Angeles, CA) with an interassay variation of 7.5% for both and an intraassay variation of 5.2% for both. Plasma postheparin extrahepatic and hepatic lipoprotein lipase levels were measured by the method of Krauss et alM after intravenous administration of 100 U of heparin per kilogram body weight, obtaining a postheparin blood sample 15 minutes later. StatisticalAnalysis The effects of exercise on lipids, lipoproteins, apolipoproteins, lipolytic enzymes, and endogenous sex hormones were assessed comparing baseline values (preexercise), with those on exercise at days 30,60, and 90 (Table 1). These data were demonstrated to be normally distributed (Shapiro-Wilk test).” Two-way ANOVA for repeated measure? was used to assess the significance of the training effect on the dependent variables (Table 1). We focused on paired observations in four cells, with days 90, 60, and 30 compared with pretraining baseline. The training effect on cardio-

MENDOZA

ET AL

respiratory fitness was also tested by two-way ANOVA4’ by comparing duration of exercise testing at a standardized submaximal load (Table 1). Cardiorespiratory fitness was further assessed by the Astrand nomogram? which is based on pulse rate at a standardized submaximal workload. Chi-square analyse? were used to compare the distributions of pulse rates at a standardized submaximal load3’ before and at the conclusion of physical training. The relationships between changes in lipids, lipoproteins, apolipoproteins, endogenous sex hormones, and lipolytic enzymes were first assessed by univariate correlations (Tables 2, 3) and then by stepwise (Tables 4 to 6) regression analysis.4’ Of the 21 variables analyzed, comparing absolute levels at baseline versus day 90, only two were not normally distributed; change in weight and change in triglyceride (Tables 2, 3). A natural log transformation of the data normalized the distribution of the change in triglyceride and further normalized, but not fully, the distribution of change in weight. Therefore, Table 2 displays both the parametric (Pearson) and nonparametric (Spearman) univariate correlation coetlicients.“s To assess the relationships between change in HDLC. change in apo A2, and change in LDLC with changes in endogenous sex hormones, measures of training effect, and change in weight, etc, we used stepwise regression for both for the transformed and untransformed data (Tables 4 to 6). To further assess whether the outcome of the stepwise regression (Tables 4,5) represented a quirk of distribution of the variables, we examined whether percent change of the variables from baseline (Table 6), which adjusts for discontinuity in the distribution of variables at baseline, would give us the same pattern of relationships as we had observed using absolute numerical change. We calculated percent change (values at day 90 minus baseline, divided by baseline). For the percent change data, of the 21 “percent change” variables, only three (percent change total cholesterol, weight, and prolactin), were not normally distributed; hence, Table 3 displays both parametric and nonparametric univariate correlation coefficients. After natural log transformation, only log percent change in total cholesterol was not normally distributed (Table 6). Considering the relatively low significance of the univariate correlation coefficients (Tables 2, 3), we also carefully assessed how much of the variance of the dependent variables we could explain in a stepwise regression if we restricted the number of independent variables that were allowed to significantly relate with the dependent variable to two or less, using the maximum R* option of the stepwise regression program (Tables 5, 6). By presenting stepwise regression models without (Table 4) and with (Tables 5. 6) restriction of the number of independent variables to two, we could examine the stability of the multiple regression coefficients and provide a variety of regression equations for cross-validation with other study cohorts. RESULTS

Effects of Training on Measures of Cardiovascular Fitness Comparing baseline (pretraining) and day 90 (end of physical training), the total time of exercise expended to exhaustion, at a standardized submaximal workload was significantly increased from 10.8 * 2 to 13.6 f 2 minutes, P I .OOl (Table 1). Separately, we compared baseline and day 90 pulse rate measured at a standardized submaximal workload.‘” By Astrand’s criteria,38 pre-exercise performance at a standardized submaximal workload was graded as average for eight patients, good for eight, and excellent for one, while at day 90, three patients were average, nine were good, and five

TRAINING, LIPOPROTEINS, SEX HORMONES

371

Table 2. Relationship of Changes in HDLC, Apo A2, and LDLC Wiih Changes in Hepatic Lipase, Lipoprotein Lipase, Estradiol, Testosterone, Prolactin, Weight, Triglyceride, Heart Rate, and Exercise Time: Pearson’s and Spearman’s Correlations AHDLC Pearson

ALDLC

AApoA2 Spearman

PearSOn

Spearman

Pearson

Spearman

AHL iP

-.41 ,098

-.34 ,177

-.21 ,428

-.17 ,510

.52

.21

,032

.43

.07

.05

,792

,647

ALPL r P A& r P

-.02

-.06

-.42

-.42

,941

,827

,090

,093

.25

.30

.36

.52

,335

,236

,162

,031

-.41

-.43

,102

,082

AT r P APRL r P AWeight

-.17 520

-.20 ,447

.30

.16

,235

,532

.07

.04

.18

.16

,803

,864

,498

,544

-.09 ,719

-.19 ,455

-.24

-.28

,362

I

.26

.25

.lO

.ll

.07

P

,317

,341

.689

.668

.791

,276 -.09 ,744

ATG r P AHR r

- .43 ,087 -.34

-.40 ,109

- .Ol ,963

-0.25

,279

,340

.37

.4%

,569

,148

,050

.06

.02

.03

.25

.23

,833

,946

,913

,336

,372

,061

I

.07

P

,790

-.lO

-.15

-.2a

,701

-.46

,177

P AExercise time

.06 ,813

Abbreviations: HL, hepatic lipase; LPL, lipoprotein lipase; E, estradiol; T, testosterone; PRL, prolactin; TG, triglyceride; HR, heart rate,

were excellent (x’ = 5, P = ,082). Hence, by Astrand’s criteria, there was a marginally significant shift toward improved aerobic cardiorespiratory fitness3’ After completing 90 days of the training period, there was also a reduction in the maximal heart rate achieved during the exercise electrocardiogram (169 v 164 bpm, P = .ll) (Table 1). Altogether, a significant training effect was realized, with an increased ability to sustain prolonged submaximal exercise. Effects of Exercise on Lipids, Lipoproteins, Apolipoproteins, Lipolytic Enzymes, and Endogenous Stx Hormones

At baseline, before initiating the exercise training, the group of men had, compared with levels for normal Venezuelan men,46 high mean triglyceride (311 mg/dL) and low HDLC levels (30 mg/dL) (Table 1). Mean pretraining E, (50.1 2 15 pg/ml) was higher than current reference intervals for men greater than 18 years old (15 to 40 pg/mL)>’ about twice as high as in normal nonsmoking men,‘5 and higher than in healthy men from Framingham (33 lr 7 pg/mL) .4K After 30 days’ exercise, there were significant decrements in triglyceride, TC and LDL cholesterol, and TC/HDLC (Table 1). After 60 days’ exercise, triglyceride, TC and LDL cholesterol, TC/HDLC, and apo B were reduced (Table 1). After 90 days’ exercise, triglyceride (P I .05) and TC/ HDLC (P I .Ol) remained lower than baseline.

Mean HDLC was increased after 30 (I’ I .05), 60 (P I .OOl), and 90 (P I .OOl) days’ exercise (Table 1). Apo A2 first manifested a significant increase after day 60, and remained elevated at day 90 (Table 1). Hepatic lipase was reduced after 30 (P I .Ol) and 90 (P I .05) days’ exercise, without changes in lipoprotein lipase (Table 1). Of the endogenous sex hormones, the major change was in E,, which decreased from an elevated baseline level of 50.1 pg/mL to 34.1,27.5, and 27.8 pg/mL, respectively, after 30, 60, and 90 days of exercise (P 5 .OOl for each), with reductions of nearly 50% by days 60 and 90 (Table 1). Mean prolactin was increased at days 60 and 90 (P 2 .05) (Table 1). There were no significant changes in testosterone, iuteinizing hormone (LH), or follicle-stimulating hormone (FSH) during the study period (Table 1). There were no significant changes in mean weight throughout the study period (P > 0.1) (Table 1). Relationships Between Changes (A) in Endogenous Se_x Hormones, Lipids, Lipoprotein Cholesterols, Apolipoproteins, and Lipolytic Enzymes Univariate correlations. Univariate relationships between changes in HDLC, changes in apo A2, and changes in LDLC were examined both for absolute change (Table 2) and percent change (Table 3). Overall, a uniform pattern of univariate and multivariate predictors for AHDLC were observed. There were three consistent explanatory vari-

372

MENDOZA

Table 3. Relationship of Percent Testosterone.

Changes in HDLC,

ET AL

Apo A2, and LDLC With Changes in Hepatic Lipase, Lipoprotein Lipase, Estradiol,

prolactin. Weight, Triglyceride, Heart Rate, and Exercise Time: Pearson’s and Spearman’s Correlations

AHDLC PearSOn

ALDLC

AApoA2 Spearman

PL?arSJll

spearman

Pearson

Spearman

AHL r

- .43

-.50

,065

P

-.25

,043

-.30

,336

,239

.35

.I7

,174

,516

ALPL r

.07

P

,600

AE, r

,970

.Ol

-.04

,961

P

-.31

-.Ol

,661

-.29

,223

,264

.54

.56

,025

,016

-.04

-.03

,662

,903

-.25

-.21

,339

,419

AT r

-.I0

-.21

,691

P

.422

.03

.14

.lQ

.05

,697

,563

.461

.a44

APRL i-

-.14

-.12

-.lQ

-.27

-.29

-.31

,956

,652

,473

,303

.253

r

.23

.25

.ll

.15

.13

P

.364

,330

,660

,566

,630

P

,219

AWeight -.05 ,644

ATG r

-.60

P

-.53

AHR r

-.I5

- .05

,661

-.36

,570

P

-.04

.027

,010

-.14

,136

-.31

,652 -.12

,564

.660

-.26

220

,305

.44

.56

,079

,020

AExercise time r

-.15

Abbreviations

-.16

,577

P

-.I1

,541

-.03

,663

,910

.27

.26

,300

,277

as in Table 2.

ables for AHDLC: Ahepatic lipase, Atriglyceride, and Aheart rate at a standardized, submaximal load (Tables 2, 3). The greater the decrease in hepatic lipase, the greater the increase in HDLC. The greater the decrease in triglyceride, the greater the increase in HDLC. The greater the decrease in heart rate, the greater the increase in HDLC.

These three variables were also significant independent predictors of AHDLC in stepwise multiple regression models (Tables 4 to 6). From Tables 2 and 3, partial RZ for AHDLC can be calculated from Atriglyceride as 19% (absolute A) or 36% (%A), from Ahepatic lipase, 17% (absolute A) or 18% (%A), and from Aheart rate, 21%

Table 4. Stepwise Regression After 3 Months’ Exercise Training: Changes in HDLC, Apo AZ, LDLC, and Changes in Triglyceride, Lipases, Endogenous Sex Steroid Hormones, Weight, Heart Rate, and Duration of Exercise Dependent Variable

ExplanatoryVariables

AHDLC AA2 AHL + ALPL + AE, + AT + APRL + AWeight + ATG + AHR + AExercise Time

ALDLC

ExplanatoryVariables

Model I72 AHDLC

.a7

F

P

11.1

.0006

F =

AHL (-) APRL

AA2

ALDLC

Abbreviations

.30

.43

as in Table 2.

3.1

5.3

.06

.OlQ

=

=

P

PartialRz

16.2

,002

4.9

.05

.14

.15

ATG (-)

37.6

.OOOl

.16

AHR (-)

16.4

,002

.33

AExercise time

5.2

.05

.04

AHL (-)

2.5

,136

.12

ALPL (-)

5.2

.036

.I6

AHL

7.3

,017

.27

AHR

3.9

.066

.I6

TRAINING, LIPOPROTEINS, SEX HORMONES

373

Table 6. Maximim t? after 3 Months’ Exercise Training: Changes in HDLC, Apo AZ, LDLC, and Changes in Triglyceride, Lipases, Endogenous Sex Steroid Hormones, Weight, Heart Rate, and Duration of Exercise Dependent Variable

Explanatory Variables

AHDLC AA2 AHL + ALPL + AE, + AT + APRL + AWeight + ATG + AHR + AExercise Time

ALDLC Best model

F

R2

AHDLC

11.8

Gi

P

F

.0004

11.0

,006

8.3

,014

AHL(-) APRL

Best l-variable

P

ATG(-)

25.5

.0003

AHR(-)

10.0

,008

model 0.18

3.4

,087

ATG(-)

3.4

,087

0.51

7.3

,007

ATG(-)

11.3

,005

AHR(-)

9.4

,008

AHL(-)

5.5

,036

ATG(-)

13.2

,003

AHR(-)

12.7

,004

3.3

.0896

Best 2-variable model

Best 3-variable

model 0.66

Best l-variable

8.3

,003

model

AA2

.I8

3.3

.0896

ALPL(-)

.43

5.3

,019

AHL

7.3

,017

AHR

3.9

,068

AH1

5.6

,032

Best model ALDLC Best l-variable

model

ALDLC

.27

5.6

,032

(absolute A) or 14% (%A). From the univariate data, from 57% to 68% of AHDLC could be accounted for by Ahepatic lipase, Atriglyceride, and Aheart rate (Tables 2, 3). Thus, the ability to predict 71% to 87% of the variance of AHDLC by multivariate analyses in Tables 4 to 6 is not surprising. For Aapo A2, there were two consistently appearing explanatory variables, Alipoprotein lipase and Aestradiol.

The greater the change in E,, the greater the change in apo A2, while the greater the change in lipoprotein lipase, the smaller the change in apo A2 (Tables 2, 3). Three variables related consistently to ALDLC: Ahepatic lipase, Aheart rate, and AE, (Tables 2, 3). The greater the decrease in hepatic lipase, the greater was the decrease in LDLC. The greater the decrease in heart rate, the greater

Table 6. Stepwise Regression After 3 Months’ Exercise Training: Changes in HDLC, A2, LDLC, and Changes in Triglyceride, Lipases, Endogenous Sex Steroid Hormones, Weight, Heart Rate, and Duration of Exercise (% A Natural Log Transformation Dependent Variable

of Variables)

Explanatory Variables

AHDLC AA2 ALDLC

AHL + ALPL + AE, + AT + APRL + AWeight + ATG + AHR + AExercise Time R* .71

Best model AHDLC

Best l-variable

-

F

P

10.5

.0009

F AHL(-)

7.0

ATG(-)

24.0

AHR(-)

P

Partial R2

.02

.I5

.0003

.44

5.4

,037

.I2

11.6

,004

model A4

11.6

.0039

ATG(-)

.59

10.0

.0020

AHL(-)

5.1

.04

ATG(-)

15.1

.002

Best P-variable model

Best model AA2

.25

5.1

,039

.38

4.2

,037

AE2

5.13

,039

.25

Best model ALDLC Best l-variable

AHL

4.5

,053

.I9

AHR

4.1

,061

.18

AHL

3.6

,078

model .19

3.6

,078

374

was the decrease in LDLC. The greater the decrease in E,, the smaller the decrease in LDLC. Multivariate correlations, AHDLC. In the stepwise multiple regression model without restriction of the number of independent variables allowed to enter the model, a considerable proportion of the variance for the increase in HDLC from baseline to day 90 (87%, P = .0006), was accounted for independently by a decrease in hepatic lipase (F = 16.2, P = .002), by a decrease in triglyceride (F = 37.6, P = .OOOl), and by a decrease in heart rate (F = 18.4, P = .002) (Table 4). Moreover, the greater the increase in exercise time (at a fixed, submaximal load), the greater was the increase in HDLC (F = 5.2, P = .OS). Restricting the number of independent variables for AHDLC allowed to enter the model to two still provided a substantial reduction of variance of AHDLC of 51% (Table 5). With AHDLC as the dependent variable, the best one-variable model included Atriglyceride (R* = IS%), the best two-variable model also included Aheart rate (R2 = 51%) and the best three-variable model included Ahepatic lipase (R* = 66%) (Table 5). To determine whether the outcome of the original stepwise analysis (Tables 4, 5) represented a quirk of distribution of the independent variables, we then examined whether percent change of the variables from baseline (Table 6), which adjusts for discontinuity in the distribution of variables at baseline, would give us the same pattern of relationships as we had observed using absolute numerical change. Using percent change (Table 6), we obtained the same pattern of significant explanatory variables (Ahepatic lipase, Atriglyceride, Aheart rate) for A% HDLC and a comparable R* (71% v 87%) as we had originally obtained from the stepwise regression model using absolute values (Tables 4, 5). Restricting the number of independent variables for AHDLC allowed to enter the model to two provided reduction of variance of AHDLC of 59% versus 71% for the unrestricted model (Table 6). Multivatiate correlations, ALDLC. A significant proportion of the variance of the decrease in LDLC could be accounted for by two independent variables at day 90 (43%, P = ,019) (Table 4). The greater the decrease in hepatic lipase, the greater was the decrease in LDLC (F = 7.27, P = .017), and the greater the decrease in heart rate, the greater was the decrease in LDLC (F = 3.90, P = .068) (Table 4). For ALDLC, the best one-variable model was Ahepatic lipase (R’ = 27%). The best two-variable model was for Ahepatic lipase and Aheart rate (Table 5), the same two variables selected in the initial stepwise regression (Table 4). In the percent change model (Table h), comparable to the absolute change (Tables 4, 5) 38% of ALDLC could be explained, again by Ahepatic lipase and Aheart rate. Multivatiate correlations, Aapo A2. For apo A2, in the original stepwise analysis (Table 4), only 30% of the variance could be explained (by two independent variables) and the model was only marginally significant (P = .08). ALPL was inversely associated with AA2 (Tables 4, 5). For the percent change model (Table 6) 25% of Aapo A2 could be accounted for, here by AEZ (P = .04). Multivariate correlations, Ahepatic lipase. Stepwise lin-

MENDOZA

ET AL

ear regression was also used to examine for determinants of changes (A) in hepatic lipase, with Ahepatic lipase (baseline v day 90) as the dependent variable, and AE, + Atestosterone + Aprolactin + Aweight + Aheart rate + Aexercise time as explanatory variables. There were no significant independent explanatory variables for Ahepatic lipase. DISCUSSION

In this group of 17 men with premature myocardial infarction, a 3-day per week, 30-minute per session exercise program, without specific diet change, without changes in drugs, and without significant weight loss, produced a significant training effect. There was a significant increase in the total time of exercise to exhaustion, at a standardized, submaximal load, and there was a borderline significant reduction in heart rate in response to a standardized submaximal load. Although some studies have reported improvements in plasma lipids with exercise training that produced no change in body weight,49-51other investigators have shown that body weight and abdominal fat loss are significant determinants of lipoprotein changes produced by exercise training.52”J Although no change in body weight was observed in the current study after training, the possibility that fat-free mass was slightly increased while fat mass was reduced cannot be excluded as a potential change in body composition. Measurement of the proportion of abdominal fat and body composition would have been valuable in the current study because the proportion of abdominal fat is a better correlate of plasma HDLC levels than obesity per se.55-57 It has been suggested that exercise training’s improvement of cardiorespiratory fitness may involve different mechanisms than those that affect plasma lipoproteins during training?8 The training regimen led to significant reductions in TC and LDLC (after 30 and 60 days), and persistent (over 90 days) decrements in triglyceride and TC/HDLC, increments in HDLC, and increments of apo A2 Our results are congruent with those summarized by Tran and Weltman” in a meta-analysis of exercise studies where there was no change in body weight; “cholesterol, LDLC, cholesterol/ HDLC, and triglyceride levels all showed statistically significant decreases.” HDLC levels and the ratio of TC to HDLC are excellent predictors of coronary heart disease.3~JZ.Y,,M) Our findings are similar to those of Hartung et a1,6’ who also reported increased HDLC levels in postmyocardial infarction patients after completion of exercise programs. Recognizing the importance of homogeneity of study conditions for our subjects, we made every effort to control changes in exogenous factors (diet, smoking, or drugs)2”X34.35 that might effect study outcome, particularly changes in HDLC and E, levels. It is possible that a wide variation in age of the patients might also have effected the outcome of the studies. Optimally, in postmyocardial infarction, physical training studies, a homogeneous sample of patients with a narrow age range, never having smoked, and taking no P-blockers would be optimal, albeit hard to find. The increases in HDLC in the current study could be due

TRAINING,

LIPOPROTEINS,

SEX HORMONES

to direct effects of exercise, through increased insulin sensitivity, through increases in adipose tissue lipoprotein lipase,b’.6” or through reduction of hepatic Iipase.63BMIn the current study, we did not measure insulin sensitivity or adipose tissue lipoprotein lipase. Hepatic lipase decreased (P < .05) 30 days after initiation of exercise training, remained at reduced levels, and reduction of hepatic lipase was independently and significantly correlated with increments in HDLC. Here, our data are similar to that of Peltonen et alM and suggest that exercise-induced reductions of hepatic lipase may directly increase HDLC, probably through reduced HDLC catabolism. Since there were no significant changes in weight, and no significant independent associations between change in weight and change in lipids, lipoproteins, and apo A2, increments in HDLC cannot be attributed to weight loss during the study. It is, of course, possible that body composition changes (reduced fat mass, increased fat-free mass), or reduced abdominal fat5’-57could have led to the changes in part in HDLC and/or apo A2. Change in HDLC and change in hepatic lipase, heart rate, and triglyceride were related. In univariate analyses, these three variables reduced variance of change in HDLC by 57% (absolute change) and 68% (percent change), and were the first three variables to appear as significant predictors of change in HDLC in the stepwise regression model. In the maximum R’ version of the stepwise regression model, the best two-variable model incorporated change in triglyceride and heart rate, which together reduced the variance of change in HDLC by 51%. This suggested that the regression model was not overparameterized; the relatively large reduction in variance of HDLC cannot be attributed simply to overparameterization. The best three-variable model also included Ahepatic lipase, which, along with Atriglyceride and Aheart rate, reduced the variance of HDLC by 66%. Thus, the univariate and multivariate assessments of explanatory variables for AHDLC were congruent, and indicated that exerciseinduced reductions in hepatic lipase, reductions in heart rate, and reductions in triglyceride were all independently and significantly associated with increases in HDLC. Measures of increased cardiorespiratory capacity were both significantly and independently related to changes in HDLC; as heart rate on a tixed, submaximal load decreased, HDLC increased, and as exercise time increased, HDLC increased. Thus, a training effect, independent of effects of training on other variables such as E,, led to a significant increase in HDLC. The exercise training effect on HDLC in our current study (increase in HDLC of 7 mg/dL [23%]) might be considered relatively marked considering the fact that the net energy deficit associated with our exercise program was modest and perhaps less than reported in some other programs. 1J.19-ZZ.Z4-2h.52-54.h1.63-hs.711 Two factors that might speculatively account for the relatively impressive exercise training effect on HDLC were altitude and triglycerides, as follows. The physical training was carried out in Merida, Venezuela, at an altitude of 5,400 feet. Exercise to a standardized submaximal heart rate at this level is presum-

375

ably greater than at much lower altitudes: hence the stated amount of exercise is greater than appreciated in other studieq5’ which have almost universally been performed at much lower altitude levels. Our cohort also had a relatively high mean triglyceride level at baseline (311 mg/dL). Given the significant inverse association between triglyceride and HDLC,“,‘” our patients might have been expected to reflect a more marked than usual increase in HDLC than a normotriglyceridemic group, particuiarly if exercise also affected triglyceride levels and then secondarily HDLC levels.” The effects of exercise on lipids and lipoprotein levels in healthy subjects have been far from uniform. In some there have been significant increases in studies, HDLC,‘q.‘2~6’~6Z while in others,@,“’ particularly in women,‘” there have been no significant changes. The male/female difference in plasma lipoprotein response to training is partly due to the fact that sedentary women generally have a more favorable plasma lipoprotein profile than men, a sex difference that probably can be attributed to gender differences in endogenous sex steroid levels. It is possible that the lack of uniformity of exercise on lipids’Y.“~*4.5’.b’.~~~~’ reflects differences in the intensity and duration of exercise, and the degree of physical training reached. In the current study, there appeared to be a significant training effect, manifested by increased total time of exercise to exhaustion at a standardized, submaximal load, and by a reduction in heart rate in response to a standardized submaximal load. In the current study, there were significant increases in apo A2 levels during the training program, changes similar to those previously reported in athletes.b8 Higher levels of apo A2 are also associated with reduced risk of coronary heart disease.69 Congruent with findings in previous studies of coronary heart disease”.?’ as a group, our patients had significant hyperestrogenemia. Their E, levels decreased significantly after 1 month of exercise, reaching lowest mean values at the end of the study. These reductions in E, could not be attributed to systematic changes in diet, which may effect EZ levels within weeks of change,“4 or changes in smoking at the beginning of the study,35 or by an initiation of drugs the beginning of the study. To the extent that hyperestrogenemia may contribute to increased risk of coronary heart disease,‘7-3” reduction of E, levels by physical training may be beneficial. However, in the full stepwise model at 90 days, which included terms for physical training, there was no significant independent relationship between change in EZ and change in HDLC. Previously, we have reported correlations between EZ and apolipoprotein levels in subjects having sustained myocardial infarction, which suggest that these hormones may, in part, modulate the apolipoprotein levels.” At day 90, there was an independent significant association between an increase in HDLC and an increase in prolactin We have not been able to document previous recognition of this relationship, nor is there a well-defined physiologic explanation for how increased prolactin might raise HDLC in men, an area deserving further study. Hamalainen et al’” found no relationship of prolactin to

376

MENDOZA

HDLC, although the correlation coefficient of prolactin with apo B was relatively high, r = .32. Although there is currently no scientific rationale known to explain the prolactin results, we chose to display them to stimulate further assessment of what may well be a physiologic effect. Similar to Peltonen et al,@ we found no significant changes in testosterone levels after exercise training. Moreover, either by simple or stepwise regression, change in testosterone was not associated with change in lipids, lipoproteins, or apolipoproteins. In contrast, Frey et al’] and Aakvaag et alz6reported reduced testosterone levels in

ET AL

subjects after completing an exercise program. However, Remes et a12’found increased testosterone levels after 6 months of exercise. These discrepant results, probably similar to those for HDLC, may possibly be explained by differences in intensity and duration of exercise, or by differing effects of exercise on body composition. In aggregate, our results suggest that a modest exercise program could have significant cardiovascular benefit for men after myocardial infarction by ameliorating their hyperestrogenemia, reducing TC and LDLC, and elevating HDLC and apo A2.

REFERENCES

1. Lowering blood cholesterol to prevent heart disease. National Institute of Health Consensus Development Conference Statement, ~015, no. 7,1985 2. Report of the National Cholesterol Education Program Expert Panel on detection, evaluation, and treatment of high blood cholesterol in adults. Arch Intern Med 148:36-69,1988 3. Grundy SM, Goodman DS, Riflrind BM, et al: The place of HDL in cholesterol management. Arch Intern Med 149:505-510, 1989 4. Livshits G, Weisbort J, Meshulam N, et al: Multivariate analysis of the twenty-year follow-up of the Donolo-Tel Aviv Prospective Coronary Artery Disease Study and the usefulness of high density lipoprotein cholesterol percentage. Am J Cardiol 63:676-681,1989 5. Wilson PW, Garrison RJ, Castelli WP, et al: Prevalence of coronary heart disease in the Framingham offspring study: Role of lipoprotein cholesterol. Am J Cardiol46:649-54, 1980 6. Gordon T, Castelli WP, Hortland MC, et al: High density lipoprotein as a predictive factor against coronary heart disease. The Framingham Study. Am J Med 62:707-714,1977 7. Miller NE, Thelle DS, Forde OH, et al: The Tromoso Heart Study; High density lipoprotein and coronary heart disease: A prospective case-control study. Lancet 2:965-968,1977 8. Frick MH, Elo 0, Haapa K, et al: Helsinki Heart Study: Primary prevention trial with Gemfibrozil in middle-aged men with dyslipidemia. N Engl J Med 317:1237-1245, 1987 9. Manninen V, Elo 0, Frick MH, et al: Lipid alterations and decline in the incidence of coronary heart disease in the Helsinki Heart Study. JAMA 260:641-651,1988 10. Ordovas JM, Schaefer EJ, Salem D, et al: Apolipoprotein Al gene polymorphism is the 3’ flanking region associated with familial hypoalphalipoproteinemia. N Engl J Med 314:671-677, 1986 11. Glueck, CJ: Nonpharmacologic and pharmacologic alteration of high density lipoprotein cholesterol: Therapeutic approaches to prevention of atherosclerosis. Am Heart J 110:11071115,1985 12. Gordon D, Probstfield JL, Garrison RJ, et al: High density lipoprotein cholesterol and cardiovascular disease: Four prospective American studies. Circulation 64:1174-1184,198l 13. Heiss G, Johnson NJ, Reiland S, et al: The epidemiology of plasma high density lipoprotein cholesterol levels. Circulation 62:IV-116-IV-136,198O (suppl4) 14. Haskell WL, Taylor HL, Wood PD, et al: Strenuous physical activity, treadmill exercise test performance, and high density lipoprotein cholesterol. Circulation 62:IV-53-IV-62,198O (suppl4) 15. Glueck CJ, Taylor H, Jacobs D, et al: Plasma high density lipoprotein cholesterol: Association with measurements of body mass. Circulation 62:1V-62-IV-70,198O (suppl4) 16. Glueck CJ, Laskarzewski PM, Suchindran CW, et al: Proge-

ny’s lipid and lipoprotein levels by parental mortality. Circulation Monogr 118 73:1-51-I-62, 1986 17. Paffenbarger RS, Nale WE: Work activity and coronary heart mortality. N Engl J Med 292:545-80, 1975 18. Epstein I, Miller GI, Sofitt IW, et al: Vigorous exercise in leisure time, coronary risk factors and resting electrocardiograms in middle aged civil servants. Br Heart J 38:403-409, 1976 19. Brownell KD, Bachorik PS, Ayerle RS: Changes in plasma lipid and lipoprotein levels in men and women after a program of moderate exercise. Circulation 65:477-484, 1982 20. Enger SG, Herbjomsen K, Erikseen J, et al: High density lipoprotein (HDL) and physical activity. Stand J Clin Lab Invest 37:251-255, 1977 21. Hartung GH, Foreyt JP, Mitchel RE, et al: Relation of diet to high density lipoprotein cholesterol in middle-aged marathon runners, joggers and inactive men. N Engl J Med 302:357-361,198O 22. Streja D, Mymn D: Moderate exercise and high density lipoprotein cholesterol. JAMA 242:2190-2192,1979 23. Moll ME, Williams RS, Lester RM, et al: Cholesterol metabolism in non-obese women. Atherosclerosis 34:159-166.1979 24. Frey MAB, Doerr BM, Laubach LL, et al: Exercise does not change high density lipoprotein cholesterol in women after 10 weeks training. Metabolism 31:1142-1146, 1982 25. Remes K, Kuoppasalmi K, Aldercreutz H: Effect of long term physical training on plasma testosterone, androstenedione, luteinizing hormone and sex-hormone-binding globulin capacity. Scan J Clin Lab Invest 39:743-749,1979 26. Aakvaag A, Sand T, Opstad PK, et al: Hormonal changes in serum in young men during prolonged physical strain. Eur J Appl Physiol39:283-291, 1978 27. Mendoza SG, Zerpa A, Carrasco H, et al: Estradiol, testosterone, apolipoproteins, lipoprotein cholesterol and lipolytic enzymes in men with premature myocardial infarction and angiographically assessed coronary occlusion. Artery 12:1-23,1983 28. Godsland IF, Wynn V, Crook D, et al: Sex, plasma lipoproteins, and atherosclerosis: Prevailing assumptions and outstanding questions. Am Heart J 114:1467-1503,1987 29. Phillips GB: Relationship between serum sex hormones and glucose, insulin and lipid abnormalities in men with myocardial infarction. Proc Nat1 Acad Sci USA 72:1729-1733, 1977 30. Entrican JH, Beach D. Carrol ACF, et al: Raised plasma estradiol and estrogen levels in young survivors of myocardial infarction. Lancet 2:487-489, 1978 31. Nordoy A, Aakraag A, Thelle D: Sex hormones and high density lipoproteins in healthy males. Atherosclerosis 34:431-436, 1979 32. Gutai J, LaPorte R, Kuller L, et al: Plasma testosterone, high density lipoprotein cholesterol and others lipoprotein fractions. Am J Cardiol48:897-902,198l 33. Mendoza SG, Osuna A, Zerpa A, et al: Hypertriglyceri-

TRAINING,

LIPOPROTEINS,

SEX HORMONES

demia and hypoalphalipoproteinemia in azoospermic and oligospermic young men: Relationships of endogenous testosterone to triglyceride and high density lipoprotein cholesterol metabolism. Metabolism 30:481-486,198l 34. Rosenthal MB, Barnard J. Rose DP, et al: Effects of high-complex-carbohydrate, low-fat, low-cholesterol diet on levels of serum lipids and estradiol. Am J Med 78:23-27,1985 35. Klaiber EL, Broverman DM, Dalen JE: Serum estradiol levels in male cigarette smokers. Am J Med 77:858-862,1984 36. Judkins M: Selective coronary arteriography. Radiology 89:81_5-. 1977 37. Bruce RA. Hornstein TR: Exercise stress testing in evaluation of patients with ischemic heart disease. Prog Cardiovasc Dis 11:371-, 1968 38. Astrand PO, Rylonint JA: A nomogram for calculation of aerobic capacity (physical fitness) from pulse rate using submaximal work. J Appl Physiol7:218-230, 1954 39. Ishikawa T, Brazier J, Steiner P, et al: A study of the heparin manganese chloride methods for determination of plasma alphalipoprotein cholesterol concentration. Lipids 11:628-639. 1976 40. Friedewald W, Levy R, Frederickson D: Estimation of the concentration of low-density lipoprotein cholesterol in plasma without use of the preparate ultracentrifuge. Clin Chem 18:499515.1972 41. Kashyap ML, Hynd B, Robinson K: A rapid and a simple method for measurement of total protein in very low density lipoprotein by the Lowry assay. J Lipid Res 21:491-494,198O 42. Cheung MC. Albers JJ: The measurement of the apolipoprotein Al and A2 levels in men and women by immunoassay. J Clin Invest 60:43-58, 1977 43. Laurel1 CB: Quantitation estimation of protein by electrophoresisin agarose gel containing antibodies. Anal Biochem 15:4552, 1966 44. Krauss AM, Levy RI, Fredrickson D: Selective measurement of two lipases activities in post-heparin plasma from normal subjects and patients with hyperlipoproteinemia. J Clin Invest 54:1107-l 124, 1974 45. SAS Procedures Guide. Cary, NC, SAS Institute, p 413: SASSTAT Guide, 1988, p 530, pp 941-947 46. Mendoza S, Nucete H, Ineichen E, et al: Lipids and lipoproteins in subjects at 3,500 and 1,000 meters altitude. Arch Environ Health 34:308-311, 1979 47. Young DS: Implementation of SI units for clinical laboratory data. Ann Intern Med 106:104-129, 1987 48. Phillips GB, Castelli WP. Abbott RD, et al: Association of hyperestrogenemia and coronary heart disease in men in the Framingham cohort. Am J Med 74:863-869,1983 49. Altekruse EG. Wilmore JH: Change in blood chemistries following a controlled exercise program. J Occup Med 15:110-l 13, 1973 50. Weltman A. Matter A, Stamford BA: Caloric restriction and/or mild exercise: Effects on serum lipids and body composition. Am J Clin Nutr 33:1002-1009, 1980 51. Tran ZV, Weltman A: Differential effects of exercise on serum lipid and lipoprotein levels seen with changes in body weight. JAMA 254:919-924, 1985 52. Wood PD, Stefanick ML, Dreon DM, et al: Changes in plasma lipids and lipoproteins in overweight men during weight loss through dieting as compared with exercise. N Engl J Med 319:1173-1179, 1988 53. Williams PT, Krauss RM, Vranizan KM, et al: Effects of exercise-induced weight loss on low density lipoprotein subfractions in healthy men. Arteriosclerosis. 9:623-632, 1989

377

54. Despres JP, Moorjani S, Tremblay A, et al: Heredity and changes in plasma lipids and lipoproteins after short-term exercise training in men. Arteriosclerosis 8:402-409, 1988 55. Haffner S, Stern MP, Hazvda HP, et al: Do upper-body and centralized adiposity measure different aspects of regional body-fat distribution? Diabetes 36:43-51, 1987 56. Despres JP. Tremblay A, Perusse L, et al: Abdominal adipose tissue and serum HDL-cholesterol: Association independent from obesity and serum triglyceride concentration. Int J Obesity 12:1-13, 1988 57. Terry RB, Wood PD, Haskell WK. et al: Regional adiposity patterns in relation to lipids, lipoprotein cholesterol, and lipoprotein subfraction mass in men. J Clin Endocrinol Metabol 68:191199,1989 58. Wood PD, Williams PT, Haskell WL: Physical activity and high density lipoproteins, in Miller NE, Miller GJ (eds): Clinical and Metabolic Aspects of High Density Lipoproteins. Amsterdam, The Netherlands, Elsevier Science, 1984, pp 133-165 59. Castelli WP, Garrison RJ, Wilson PWF. et al: Incidence of coronary heart disease and lipoprotein cholesterol levels. JAMA 256:2835-2838,1986 60. Goldbourt V, Medalie JH: High density lipoprotein cholesterol and incidence of coronary heart disease. The Israeli Ischemic Heart Disease Study. Am J Epidemiol 109:296-308, 1979 61. Hartung H, William GS, Gotto AM Effect of exercise training on plasma high density lipoprotein cholesterol in coronary disease patients. Am Heart J 1010:181-184.1981 62. Nikkila EA: Metabolic and endocrine control of plasma high density lipoprotein concentration relation to catabolism of triglyceride rich lipoproteins, in Gotto AM Jr, Miller NE. Oliver MF (eds): High Density Lipoproteins and Atherosclerosis. Amsterdam, The Netherlands, Elsevier-North Holland, 1978, p 177 63. Stubbe I, Honsson P, Gustafson A, et al: Plasma lipoprotein and lipolytic enzyme activities during endurance training in sedentary men: Changes in high density subfractions and composition. Metabolism 32:1120-1128,1983 64. Peltonen P, Marniemi J, Hietanen E, et al: Changes in serum lipids, lipoproteins and heparin releasable lipolytic enzymes during moderate physical training in men: A longitudinal study. Metabolism 30:518-526,198l 65. Higuchi M, Hashimoto I, Yamakawa K, et al: effect of exercise training on plasma high density lipoprotein cholesterol level at constant weight. Clin Physiol4:125-133.1984 66. Lipson LC, Bonow R, Schaefer E, et al: Effect of exercise conditioning on plasma high density lipoproteins and other lipoproteins. Atherosclerosis 37:529-538,198O 67. LaRosa JC, Cleary P, Muesing R, et al: Effect of long term moderate physical exercise of plasma lipoproteins. Arch Intern Med 142:2269-2274.1982 68. Lehtonen A, Viikari J, Ehngholm C: The effect of exercise on high density (HDL) lipoprotein, apoproteins. Acta Physiol Stand 106:487-488,1979 69. Avogaro P, Bittolo Bon G, Cazzolato G, et al: Relationship between apolipoproteins and chemical components of lipoproteins in survivors of myocardial infarction. Atherosclerosis 37:69-76, 1980 70. Hamalainen E, Adlercreutz H, Ehnholm C. et al: Relationships of Serum lipoproteins and apoproteins to sex hormones and to the binding capacity of sex hormone binding glubulin in healthy Finnish men. Metabolism 35:535-541,1986 71. Frey MB, Doerr B, Srivastava LS, et al: Exercise training, sex hormones, and lipoprotein relationships in men. Am J Appl Physiol54:757-762, 1983