Cancer Causes & Control

, Volume 18, Issue 7, pp 743–752

Physical activity and inactivity in relation to sex hormone, prolactin, and insulin-like growth factor concentrations in premenopausal women

Exercise and premenopausal hormones

Authors

    • Channing Laboratory, Department of MedicineBrigham and Women’s Hospital and Harvard Medical School
    • Department of EpidemiologyHarvard School of Public Health
  • Stacey A. Missmer
    • Channing Laboratory, Department of MedicineBrigham and Women’s Hospital and Harvard Medical School
    • Department of EpidemiologyHarvard School of Public Health
    • Department of Obstetrics, Gynecology, and Reproductive BiologyBrigham and Women’s Hospital and Harvard Medical School
  • A. Heather Eliassen
    • Channing Laboratory, Department of MedicineBrigham and Women’s Hospital and Harvard Medical School
    • Department of EpidemiologyHarvard School of Public Health
  • Robert L. Barbieri
    • Department of Obstetrics, Gynecology, and Reproductive BiologyBrigham and Women’s Hospital and Harvard Medical School
  • Mitch Dowsett
    • Academic Department of BiochemistryRoyal Marsden Hospital
  • Susan E. Hankinson
    • Channing Laboratory, Department of MedicineBrigham and Women’s Hospital and Harvard Medical School
    • Department of EpidemiologyHarvard School of Public Health
Original Paper

DOI: 10.1007/s10552-007-9017-5

Cite this article as:
Tworoger, S.S., Missmer, S.A., Eliassen, A.H. et al. Cancer Causes Control (2007) 18: 743. doi:10.1007/s10552-007-9017-5

Abstract

An association between physical activity and premenopausal breast cancer risk may be due, in part, to relationships with sex hormones or growth factors. Therefore, we assessed whether MET-h/week of total physical activity (moderate-to-vigorous intensity), walking, or vigorous physical activity, and h/week of standing or sitting were associated with plasma concentrations of several hormones. We examined levels of estrogens, androgens, progesterone, prolactin, sex hormone binding globulin (SHBG), insulin-like growth factor-1 (IGF-1), IGF binding protein-3, and growth hormone (GH) in 565 premenopausal women, ages 33–52 years, from the Nurses’ Health Study II (NHSII). About 87% of women had both timed follicular and luteal samples; other women had one untimed sample. In general we observed few associations between sex hormone or IGF levels and measures of physical activity or inactivity. However, free testosterone was modestly inversely associated with total physical activity (p-trend = 0.02). Luteal estradiol, free estradiol, and estrone also were inversely associated with total physical activity (p-trend = 0.10, 0.04, 0.01, respectively); however, the trend was substantially attenuated when excluding women with anovulatory cycles or irregular cycles. These cross-sectional results suggest that physical activity and inactivity have limited associations with premenopausal sex hormone and growth factor levels, except possibly luteal estrogens.

Keywords

Physical activityPremenopausal womenSex hormonesProlactinIGF-1

Introduction

Physical activity is associated with a decreased risk of breast cancer in postmenopausal women and possibly premenopausal women [1]. In postmenopausal women, physical activity can reduce estrogen levels [2], a potent breast cancer risk factor [3], primarily through a reduction in body fat [2]. However, less is known about the relationship of physical activity with premenopausal sex hormone and growth factor levels. Some studies suggest that vigorous physical activity may lower estrogen levels by inducing anovulatory menstrual cycles in premenopausal women [47], but it is unclear whether activity may alter hormone levels through other mechanisms.

Few studies have examined the relationship between physical activity and levels of sex and growth hormones specifically in premenopausal women. One study including 636 premenopausal women reported a modest, inverse relationship between estradiol at any time in the menstrual cycle and hours/week of vigorous exercise [8]. Further, physical activity was not associated with insulin-like growth factor-1 (IGF-1) or its binding protein (IGFBP-3) in several other studies [914]; however, activity has been associated with transient increases in prolactin and growth hormone (GH) levels [4, 1520]. A trial of resistance training for 21 weeks observed no changes in testosterone, free testosterone, DHEAS, or IGF-1 among 12 premenopausal women [15].

Therefore we examined physical activity (total, walking, or vigorous activity) and physical inactivity (standing or sitting), in relation to plasma concentrations of estrogens, androgens, progesterone, prolactin, sex hormone binding globulin (SHBG), IGF-1, IGFBP-3, and GH in nearly 600 premenopausal women, ages 33–52 years, from the Nurses’ Health Study II (NHSII). A unique aspect of this study is that timed blood samples were collected in both the early follicular and midluteal phases of the menstrual cycle, allowing the assessment of phase-specific associations for the estrogens.

Methods

Study population

The NHSII was established in 1989, when 116,609 female registered nurses, aged 25–42, completed a questionnaire. The cohort has been followed biennially since inception to update exposure variables and ascertain newly diagnosed disease. The racial/ethnic breakdown is 94% white, 2% Asian, 2% African-American, and 2% Hispanic.

Between 1996 and 1999, 29,611 cohort members, aged 32–54 years, provided a blood sample (described in [21]). Briefly, premenopausal women who had not taken hormones, been pregnant, or breastfed within six months (n = 18,521) answered a short questionnaire and provided timed blood samples on the third to fifth day of their menstrual cycle (follicular sample) and seven to nine days before the anticipated start of their next cycle (luteal sample). Follicular plasma was aliquoted by the participant 8–24 h after collection and frozen. All other women (n = 11,090) provided a single untimed blood sample. Luteal and untimed samples were shipped via overnight courier, processed by our laboratory, and separated into plasma, red blood cell, and white blood cell components. Samples have been stored in continuously monitored, liquid nitrogen freezers since collection.

We restricted the analysis to premenopausal women, who were defined as providing timed samples, or for women providing untimed samples reporting that her periods had not ceased or reporting having a hysterectomy, but with at least one ovary remaining and was ≤47 (for nonsmokers) or ≤45 (for smokers) years of age. Follow-up of the blood cohort was 98% in 2003. Participants in this study were controls from a nested case–control study who were matched to breast cancer cases diagnosed after blood collection and before June 2003 (detailed in [21]). These women were matched to breast cancer cases on age, menopausal status at blood collection and diagnosis, month/year of blood draw, ethnicity, luteal day of the blood collection (timed samples only, date of next period–date of luteal draw), and for each blood draw, time of day and fasting status. We also included a subset of women who were in a reproducibility study [22]; these women provided follicular and luteal blood samples three times over a three-year period. The study was approved by the Committee on the Use of Human Subjects in Research at the Brigham and Women’s Hospital.

Covariate data

Information on exposures and potential covariates were asked on a questionnaire completed at blood collection and the biennial study questionnaires. Oral contraceptive use, parity, age at menarche, cycle regularity between ages 18 and 22, and height were reported at baseline in 1989; oral contraceptive use and parity were updated on subsequent biennial questionnaires. Current weight and details about the blood collection date, time, and fasting status were reported on the blood questionnaire. Current BMI was calculated as weight in kg divided by height in meters squared.

In 1997, women reported their average amount of time spent per week during the previous year in each of eight activities (walking or hiking outdoors, jogging, running, bicycling, racquet sports, lap swimming, calisthenics, or other aerobic activities) and the number of flights of stairs climbed per day [23]. Walking pace also was reported (easy, average, brisk, very brisk, unknown). A metabolic equivalent score (MET) was assigned to each activity and, based on the amount of time a woman reported a particular activity, we calculated the MET-h/week of total physical activity, walking, and vigorous physical activity (jogging, running, biking, swimming, racquet sports, other aerobic activities, and stair climbing) as previously described [23]. We selected physical activity cutpoints in multiples of 3 MET-h/week, as this is equivalent to one hour of walking. We also asked women to report if they did not exercise as much during a particular season or if they had an impairment contraindicating exercise. A previous validation study of this physical activity questionnaire observed a correlation between total MET-h/week from the questionnaire and four 7-day physical activity diaries of 0.62 [24]. In addition, we asked women to report the total number of hours per week they spent standing at work or home and sitting at work, watching television, or other sitting at home. These were summed to provide the total number of hours/week standing and sitting.

Laboratory assays

Hormone assay methods for estrogens and testosterone have been described previously [25]. In brief, samples were assayed at Quest Diagnostics (San Juan Capistrano, CA) by radioimmunoassay following extraction and celite column chromatography. After extraction of estrone, enzyme hydrolysis, extraction, and column chromatography, estrone sulfate was assayed by radioimmunoassay of estrone. Free estradiol and testosterone were calculated per Sodergard et al. [26]. At the Royal Marsden Hospital, DHEA and androstenedione were assayed by radioimmunoassay (Diagnostic Systems Laboratories, TX), and DHEAS, SHBG, and progesterone were measured by chemiluminescent immunoassay using the Immulite auto-analyzer (Diagnostic Products Corp., UK). Prolactin was measured using a microparticle enzyme immunoassay at the Massachusetts General Hospital, using the AxSYM Immunoassay system (Abbott Diagnostics, Chicago, IL). Total IGF-1, IGFBP-3, and GH levels were assayed by ELISA after acid extraction, using reagents from Diagnostic Systems Laboratory (Webster, TX, USA). We measured the hormones on the following sample sets: estradiol, estrone, and estrone sulfate in follicular and luteal samples; testosterone, androstenedione, prolactin, and SHBG in follicular, luteal, and untimed samples; IGF-1, IGFBP-3, GH, DHEA and DHEAS in luteal and untimed samples; and progesterone in luteal samples.

Follicular and luteal samples from each woman were assayed together; samples were assayed in three batches. The inter-assay coefficients of variation (CVs) from masked replicate samples in each batch were 6–14% for all hormones except progesterone (CV = 17%). Correlations from a subset of 12 samples run in two of these batches were >0.90 for all hormones.

Statistical analysis

We excluded 27 women who reported having an impairment contraindicating exercise. For each analyte, we excluded women with missing values related to assay difficulties or low volume. We also identified and excluded a small number of values (n ≤ 6 per hormone) that were statistical outliers [21, 22, 2729] based on the generalized extreme studentized deviate many-outlier detection approach [30]. For the estrogens, we examined the associations with follicular and luteal measures separately. For testosterone, androstenedione, prolactin, and SHBG, we averaged the follicular and luteal values, as levels were similar between these two menstrual cycle phases [21, 22, 27, 29]. Women with untimed samples were included for testosterone, androstenedione, prolactin, SHBG, DHEA, DHEAS, GH, IGF-1, and IGFBP-3 because levels do not differ substantially across the menstrual cycle for these hormones [31].

Primary analyses calculated adjusted geometric means by category of exposure, using a general linear model. Exposures consisted of total physical activity (<3, 3 to <9, 9 to <18, 18 to <27, 27 to <42, 42+ MET-h/week of a moderate-to-vigorous intensity), walking (<3, 3 to <9, 9 to <18, 18+ MET-h/week), vigorous physical activity (<3, 3 to <9, 9 to <18, 18+ MET-h/week), total inactivity (standing plus sitting, quartiles, <55, 55 to <73, 73 to <97.5, 97.5+ h/week), standing (quartiles, <23.5, 23.5 to <46, 46 to <61, 61+ h/week), and sitting (quartiles, <19.5, 19.5 to <27, 27 to <42, 42+ h/week). We examined both physical activity and inactivity, because these measures are independently related to BMI, triglycerides, HDL cholesterol, leptin, and type 2 diabetes [3236], and in our study the correlation between total physical activity and inactivity was 0.06 (Supplemental Table A). Tests for trend were conducted by modeling continuous exposure measures and calculating the Wald statistic [37]. For all exposures we conducted secondary analyses restricting to women with ovulatory cycles (for timed samples), to women who reported having regular menstrual cycles from age 18––22 years, and to women reporting consistent exercise patterns year round. Further, we stratified analyses by BMI at blood draw and, for walking, by walking pace.

Multivariate models adjusted for assay batch (1, 2, 3), age at blood draw (<40, 40 to <45, 45+yr), follicular and luteal fasting status (≤10, >10h), follicular and luteal time of day of blood draw (1–8 a.m., 9 a.m.-noon, 1 p.m.-midnight), month of blood draw (continuous), difference between luteal blood draw date and date of the next menstrual period (continuous, untimed), follicular day (continuous, untimed), duration of past oral contraceptive use (never, <4, 4+year, missing), ovulatory status at the blood draw (ovulatory [defined as progesterone ≥400 ng/dl], anovulatory, untimed), parous (yes, no), family history of breast cancer (yes, no), menstrual regularity at age 18–22 (very regular, regular, irregular), and age at menarche (<12, 12, 13, 14+year). In secondary analyses, we additionally adjusted for BMI (continuous). In the analyses for vigorous exercise we adjusted for walking and visa versa. Analyses of inactivity were adjusted for total physical activity in MET-h/week. P-values were two-sided and considered statistically significant if ≤0.05.

Results

Among 565 women available for analysis, the mean age was 43 years (Table 1). Eighty-seven percent of women provided timed samples, and of those, 91% were ovulatory; 77% of women reported regular cycles at ages 18–22 years. Average total physical activity was 20.5 MET-h/week (median = 15.2 MET-h/week), of which walking comprised an average 6.6 MET-h/week. Mean BMI at blood collection was 24.8 kg/m2. Hormone levels were in the expected ranges for premenopausal women [38].
Table 1

Characteristics at blood draw of premenopausal women (Nurses’ Health Study II)

 

Mean (SD)/% (n = 565)

Age (year)

43.3 (3.9)

Age at menarche (year)

12.5 (1.4)

Paritya

2.3 (1.0)

BMI at blood draw (kg/m2)

24.8 (5.0)

Total physical activity (MET-h/week)

20.5 (21.8)

Walking (MET-h/week)

6.6 (8.3)

Vigorous activity (MET-h/week)

10.6 (17.0)

Standing (h/week)

44.8 (27.0)

Sitting (h/week)

31.8 (18.8)

Never used oral contraceptives, %

15.8

Timed sample, %

86.6

Ovulatory cycle, %

91.0

Regular menstrual cycles (ages 18–22), %

77.3

Family history of breast cancer, %

9.6

 

Median (10th–90th percentile)

Estradiol, pg/ml

Follicular

46 (21–98)

Luteal

123 (67–205)

Free estradiol, pg/ml

Follicular

0.6 (0.3–1.1)

Luteal

1.6 (0.9–2.6)

Estrone, pg/ml

Follicular

40 (25–63)

Luteal

79 (48–124)

Estrone sulfate, pg/ml

Follicular

668 (303–1,548)

Luteal

1,441 (552–3,235)

Testosterone, ng/dl

24 (15–37)

Free testosterone, ng/dl

0.19 (0.11–0.35)

Androstenedione, ng/dl

106 (62–174)

DHEAb ng/dl

639 (341–1,159)

DHEAS, ug/dl

80 (40–147)

Progesterone, ng/dl

1,417 (361–2,547)

Prolactin, ng/ml

15 (8–26)

SHBG, nmol/l

64 (33–109)

GH, ng/ml

0.25 (0.14–4.15)

IGF-1, ng/ml

245 (147–348)

IGFBP-3, ng/ml

4,760 (3,235–5,896)

aAmong parous women only

bAbbreviations: DHEA = dehydroepiandrosterone; DHEAS = dehydroepiandrosterone sulfate; SHBG = sex hormone binding globulin, IGF = insulin-like growth factor, BP = binding protein, and GH = growth hormone

Total moderate-to-vigorous intensity physical activity levels were inversely associated with free testosterone (p-trend = 0.02), luteal free estradiol (p-trend = 0.04), and luteal estrone levels (p-trend = 0.01), and were marginally inversely associated with luteal estradiol levels (p-trend = 0.10) (Table 2). Compared to women exercising <3 MET-h/week, those exercising 42+ MET-h/week had 10% lower free testosterone, 14% lower luteal free estradiol, 10% lower luteal estrone levels, and 9% lower luteal estradiol levels. The results were similar when adjusting for current BMI (Supplemental Table B); however, the trends for the estrogens were substantially attenuated when excluding women with anovulatory cycles and irregular menstrual cycles (Supplemental Table C). For example, women exercising 42+ MET-h/week had an 8% lower luteal free estradiol (p-trend = 0.44), 2% lower luteal estrone (p-trend = 0.45), and 5% lower luteal estradiol (p-trend = 0.59) versus women exercising < 3 MET-h/week. In general we did not observe associations with the other hormones in primary (p-trends ≥ 0.13) or secondary analyses (p-trends ≥ 0.11).
Table 2

Adjusted geometric mean levelsa of sex and growth hormones by total physical activity in premenopausal women

 

Total physical activity (MET-h/week)

n

<3

3 to <9

9 to <18

18 to <27

27 to <42

42+

p-trendb

Sample size range

 

59–76

80–116

100–132

62–76

60–75

57–72

 

Estradiol, pg/ml

Follicular

426

44

50

49

46

44

49

0.96

Luteal

423

123

121

120

119

112

112

0.10

Free estradiol, pg/ml

Follicular

404

0.57

0.66

0.59

0.56

0.55

0.62

0.83

Luteal

415

1.65

1.56

1.56

1.47

1.44

1.42

0.04

Estrone, pg/ml

Follicular

429

38

42

41

44

37

38

0.53

Luteal

459

80

80

79

80

68

72

0.01

Estrone sulfate, pg/ml

Follicular

419

612

751

716

699

574

698

0.72

Luteal

418

1,392

1,446

1,390

1,357

1,191

1,481

0.51

Testosterone, ng/dl

511

22

24

24

26

23

22

0.20

Free testosterone, ng/dl

507

0.20

0.20

0.19

0.20

0.18

0.18

0.02

Androstenedione, ng/dl

518

95

108

104

115

100

100

0.74

DHEAc, ng/dl

433

593

663

639

731

592

608

0.33

DHEAS, μg/dl

432

72

82

76

89

73

72

0.22

Progesterone, ng/dl

462

1,214

1,203

1,105

1,178

1,134

1,112

0.24

SHBG, nmol/l

531

56

59

63

67

63

62

0.13

Prolactin, ng/ml

524

15

16

14

16

15

14

0.46

GH, ng/ml

437

0.46

0.47

0.47

0.58

0.50

0.54

0.31

IGF-1, ng/ml

543

226

228

241

233

236

230

0.82

IGFBP-3, ng/ml

543

4,619

4,545

4,558

4,615

4,433

4,603

0.98

aAdjusted for assay batch, age at blood draw, fasting status (follicular and luteal phase), time of day of blood draw (follicular and luteal phase), month of blood draw, luteal difference, follicular day, timing, duration of oral contraceptive use, family history of breast cancer, menstrual regularity, parity, and ovulatory status

bTrend across continuous total activity in MET-h/week, using the Wald test

cAbbreviations: DHEA = dehydroepiandrosterone; DHEAS = dehydroepiandrosterone sulfate; SHBG = sex hormone binding globulin, IGF = insulin-like growth factor, BP = binding protein, and GH = growth hormone

Walking was modestly and inversely associated with IGF-1 levels (p-trend = 0.01), with 6% lower levels among women walking 18+ versus < 3 MET-h/week (Table 3). This result remained after adjustment for BMI (p-trend = 0.01). We did not observe any other significant associations between walking and the hormone outcomes in primary analyses (p-trends ≥ 0.08), after adjusting for BMI (p-trends ≥ 0.07), when excluding women with anovulatory cycles or irregular cycles (p-trends ≥ 0.18), or among women who exercised consistently year-round (p-trends ≥ 0.12). Walking pace was not significantly associated with any hormone (data not shown).
Table 3

Adjusted geometric mean levelsa of sex and growth hormones by walking in premenopausal women

 

Walking (MET-h/week)

n

<3

3 to  <9

9 to <18

18+

p-trendb

Sample size range

 

160–210

147–191

66–88

45–58

 

Estradiol, pg/ml

Follicular

426

49

47

42

51

0.99

Luteal

423

116

121

125

112

0.64

Free estradiol, pg/ml

Follicular

404

0.60

0.60

0.51

0.67

0.82

Luteal

415

1.46

1.56

1.64

1.44

0.65

Estrone, pg/ml

Follicular

429

40

41

40

39

0.51

Luteal

459

76

79

81

68

0.17

Estrone sulfate, pg/ml

Follicular

419

661

716

658

674

0.56

Luteal

418

1,303

1,466

1,444

1,267

0.64

Testosterone, ng/dl

511

24

24

24

22

0.25

Free testosterone, ng/dl

507

0.19

0.20

0.19

0.18

0.27

Androstenedione, ng/dl

518

104

105

107

96

0.31

DHEAc, ng/dl

433

626

665

652

570

0.18

DHEAS, μg/dl

432

75

83

77

69

0.17

Progesterone, ng/dl

462

1,209

1,098

1,227

1,062

0.14

SHBG, nmol/l

531

63

60

62

61

0.80

Prolactin, ng/ml

524

15

15

15

14

0.10

GH, ng/ml

437

0.41

0.60

0.52

0.49

0.33

IGF-1, ng/ml

543

236

236

227

221

0.01

IGFBP-3, ng/ml

543

4,574

4,589

4,521

4,466

0.08

aAdjusted for assay batch, age at blood draw, fasting status (follicular and luteal phase), time of day of blood draw (follicular and luteal phase), month of blood draw, luteal difference, follicular day, timing, duration of oral contraceptive use, family history of breast cancer, menstrual regularity, parity, vigorous physical activity (MET-h/week), and ovulatory status

bTrend across continuous walking in MET-h/week, using the Wald test

cAbbreviations: DHEA = dehydroepiandrosterone; DHEAS = dehydroepiandrosterone sulfate; SHBG = sex hormone binding globulin, IGF = insulin-like growth factor, BP = binding protein, and GH = growth hormone

Overall we did not observe any clear associations between vigorous physical activity and hormone concentrations in premenopausal women in primary (p-trends ≥ 0.09) (Table 4) or secondary analyses (p-trends, adjusted for BMI ≥0.15; p-trends, exclude women with anovulatory cycles or irregular cycles ≥ 0.09). However, after adjustment for BMI, luteal estrone was significantly inversely associated with vigorous physical activity (p-trend = 0.05), with 10% lower levels among women vigorously exercising 18+ versus < 3 MET-h/week. This trend was substantially attenuated after excluding women with anovulatory cycles or irregular menstrual cycles (p-trend = 0.67).
Table 4

Adjusted geometric mean levelsa of sex and growth hormones by vigorous activity in premenopausal womenb

 

Vigorous activity (MET-h/week)

n

<3

3 to <9

9 to <18

18+

p-trendb

Sample size range

 

162–220

108–141

64–85

84–101

 

Estradiol, pg/ml

Follicular

426

46

49

49

46

0.79

Luteal

423

122

120

111

116

0.25

Free estradiol, pg/ml

Follicular

404

0.57

0.64

0.58

0.59

0.79

Luteal

415

1.55

1.62

1.35

1.49

0.19

Estrone, pg/ml

Follicular

429

40

42

39

40

0.88

Luteal

459

81

77

69

75

0.09

Estrone sulfate, pg/ml

Follicular

419

687

691

624

701

0.64

Luteal

418

1,425

1,383

1,186

1,442

0.70

Testosterone, ng/dl

511

24

23

22

24

0.64

Free testosterone, ng/dl

507

0.20

0.19

0.17

0.20

0.21

Androstenedione, ng/dL

518

105

101

101

109

0.73

DHEAc, ng/dl

433

639

620

633

660

0.87

DHEAS, μg/dl

432

81

75

72

77

0.39

Progesterone, ng/dl

462

1,149

1,214

1,139

1,102

0.51

SHBG, nmol/l

531

62

58

68

60

0.34

Prolactin, ng/ml

524

15

15

14

15

0.58

GH, ng/ml

437

0.50

0.46

0.58

0.49

0.52

IGF-1, ng/ml

543

225

246

233

232

0.16

IGFBP-3, ng/ml

543

4,530

4,622

4,504

4,585

0.38

aAdjusted for assay batch, age at blood draw, fasting status (follicular and luteal phase), time of day of blood draw (follicular and luteal phase), month of blood draw, luteal difference, follicular day, timing, duration of oral contraceptive use, family history of breast cancer, menstrual regularity, parity, walking (MET-h/week), and ovulatory status

bTrend across continuous vigorous activity in MET-h/week, using the Wald test

cAbbreviations: DHEA = dehydroepiandrosterone; DHEAS = dehydroepiandrosterone sulfate; SHBG = sex hormone binding globulin, IGF = insulin-like growth factor, BP = binding protein, and GH = growth hormone

We observed few associations between h/week of sitting (p-trends ≥ 0.10) (Table 5), standing (p-trends ≥ 0.14), or total physical inactivity (standing plus sitting, p-trends ≥ 0.11) and hormone levels (data not shown). GH was inversely associated with hours of inactivity (% difference, top versus bottom quartile = -24%, p-trend = 0.05). This same association was observed with hours per week of standing (p-trend = 0.04), but not with sitting (p-trend = 0.61). Overall there were few interactions between total physical activity and hours per week of sitting (p ≥ 0.11 after excluding women with anovulatory cycles or irregular menstrual cycles, data not shown), with the exception of SHBG (p-interaction = 0.02). This potential interaction was attenuated after adjusting for BMI (p-interaction = 0.17).
Table 5

Adjusted geometric mean levelsa of sex and growth hormones by hours per week of sitting among premenopausal women

 

Sitting (h/week)

n

<19.5

19.5 to <27

27 to <42

42+

p-trendb

Sample size range

 

107–129

75–101

118–160

105–139

 

Estradiol, pg/ml

Follicular

412

46

49

45

50

0.57

Luteal

409

118

113

123

118

0.93

Free estradiol, pg/ml

Follicular

391

0.58

0.62

0.56

0.63

0.60

Luteal

401

1.52

1.49

1.60

1.48

0.64

Estrone, pg/ml

Follicular

415

38

38

41

43

0.10

Luteal

444

76

76

78

77

0.99

Estrone sulfate, pg/ml

Follicular

406

712

646

674

707

0.75

Luteal

403

1,368

1,244

1,459

1,404

0.63

Testosterone, ng/dl

495

24

22

23

24

0.42

Free testosterone, ng/dl

491

0.19

0.19

0.19

0.20

0.99

Androstenedione, ng/dl

501

105

95

104

111

0.19

DHEAc, ng/dl

419

628

632

642

639

0.73

DHEAS, μg/dl

418

77

74

79

77

0.56

Progesterone, ng/dl

447

1,208

1,122

1,174

1,126

0.62

SHBG, nmol/l

513

61

59

63

63

0.46

Prolactin, ng/ml

506

15

15

15

16

0.18

GH, ng/ml

423

0.48

0.56

0.55

0.43

0.61

IGF-1, ng/ml

525

234

242

229

234

0.65

IGFBP-3, ng/ml

525

4,491

4,711

4,554

4,549

0.87

aAdjusted for assay batch, age at blood draw, fasting status (follicular and luteal phase), time of day of blood draw (follicular and luteal phase), month of blood draw, luteal difference, follicular day, timing, duration of oral contraceptive use, family history of breast cancer, menstrual regularity, parity, and ovulatory status

bTrend across continuous sitting in h/week, using the Wald test

cAbbreviations: DHEA = dehydroepiandrosterone; DHEAS = dehydroepiandrosterone sulfate; SHBG = sex hormone binding globulin, IGF = insulin-like growth factor, BP = binding protein, and GH = growth hormones

Discussion

We observed that measures of physical activity and inactivity were not strongly associated with mid- to late-premenopausal sex hormone or growth factor concentrations. Total physical activity was inversely associated with luteal estrogen levels overall, although the association was attenuated after excluding women with anovulatory cycles or irregular cycles. This suggests that high levels of physical activity (≥42 MET-h/week, equivalent to seven hours per week of vigorous physical activity), by inducing anovulatory or irregular menstrual cycles [5, 3840], suppress luteal estrogen levels. Our data are consistent with a previous study in premenopausal women (n = 636), which observed lower estradiol levels across the entire menstrual cycle in active women [8]; another study (n = 50) observed no association between current physical activity and luteal or follicular estradiol, perhaps due to the small sample size [41]. Interestingly, we and Verkasalo et al. [8] did not observe a significant inverse trend in progesterone levels. However, in our study, women exercising ≥42 versus < 3 MET-h/week had 8% lower progesterone levels, providing some support for the anovulation hypothesis. Despite the apparent reductions in luteal estradiol with increased physical activity, previous prospective studies have not observed an association, to date, between luteal estrogens and premenopausal breast cancer risk [27, 42]. This suggests that anovulation may be only one mechanism through which physical activity influences risk among women with moderate amounts of physical activity. Consistent with a previous large study [8], generally we did not observe associations between other sex hormones or SHBG and physical activity or inactivity. Although BMI has been associated with sex hormones in premenopausal women [8, 43], the correlation between BMI and total physical activity was only −0.14 in our study (Supplemental Table A).

We also did not observe associations between prolactin or GH and physical activity. Previous literature suggests that both prolactin and growth hormone are increased after acute exercise [4, 1620] or resistance training [15]. Both hormones, however, decrease to baseline levels within 24 h of the exercise bout [44]. Our results suggest that increased exercise is not associated with long-term increases in prolactin or growth hormone levels. This is consistent with one randomized, yearlong trial of physical activity, which did not observe an effect of the intervention on prolactin levels [45]; to our knowledge no long-term studies of GH are available. Although GH was inversely associated with h/week of inactivity, given little biologic plausibility, this association should be interpreted with caution.

Overall we also did not observe associations between physical activity or inactivity and IGF-1 or IGFBP-3. IGF-1 was associated with walking; however, given that we did not observe an association with total or vigorous physical activity, this result may be due to chance. Our results are consistent with several previous studies in adolescent girls and premenopausal women, which observed little or no association between various measures of physical activity and IGF-1 or IGFBP-3 [9, 1114, 46]. In total these results strongly suggest that physical activity is not importantly related to these biomarkers.

Physical activity has been hypothesized to influence breast cancer risk via hormonal mechanisms or growth factors [47]—in general, though, our results do not support this hypothesis. Interestingly, two recent case–control studies have reported that physical activity was similarly inversely associated with both ER+ and ER- tumors among premenopausal women [48, 49], suggesting that exercise may act through nonhormonal or other mechanisms in this population. Animal evidence also is consistent with a nonhormonal hypothesis. Among 160 rats in a randomized moderate exercise intervention study, estradiol and progesterone levels did not differ across the 8-week intervention between exercising and control rats [50]. Interestingly, mammary apoptosis was increased in the exercise group versus controls [50]. It is possible that this mechanism may translate into human populations.

The primary limitation of this study is its cross-sectional nature, precluding the ability to determine causal relationships. However, this is a large study, with nearly 600 premenopausal women, most of whom have carefully timed follicular and luteal blood samples. This collection is unique to our study, allowing phase-specific analyses for estrogens and reducing confounding by time in the menstrual cycle. A second limitation is that a single blood sample provides a somewhat imprecise measure of long-term average hormone levels [22], and thus could attenuate our results. Additionally, both prolactin and DHEA are affected by circadian rhythms; however results were similar after restricting to blood draws conducted before 11 a.m. Physical activity as measured by this questionnaire had a reasonable correlation with physical activity diaries (r = 0.62), thus our results may be slightly attenuated. However, physical activity previously has been associated with gestational diabetes [51], ovulatory infertility [52], and possibly breast cancer risk [23] in the same cohort, suggesting that the lack of an observed association with premenopausal sex hormones is not likely to be the result of exposure misclassification. The median level of total physical activity in our population, equivalent to about five hours per week of moderate intensity physical activity, is somewhat higher than in the general population. However there was still a wide spread of total physical activity levels in our participants allowing us to examine the relationship between this and sex hormone levels.

In conclusion, our results suggest that physical activity and inactivity do not have clear associations with adult premenopausal hormone levels, with the exception of reductions in luteal estrogen levels. These results should be confirmed in other studies of premenopausal women with carefully timed, and perhaps serial, blood samples. However, other potential mechanisms through which physical activity may influence premenopausal breast cancer risk should be explored.

Supplementary material

Copyright information

© Springer Science + Business Media B.V. 2007