No effect of exercise on insulin-like growth factor (IGF)-1, insulin and glucose in young women participating in a 16-week randomized controlled trial

Abstract

Introduction

Insulin-like growth factor 1 (IGF-1) and IGF binding protein 3 (IGFBP-3) have been associated with increased risk of breast cancer.

Methods

We report our findings on the effects of 16 weeks of aerobic exercise on IGF axis proteins, insulin, glucose and insulin resistance of 319 young sedentary women. Demographics, health surveys, body composition, dietary intake and blood samples were collected at baseline and 16 weeks. IGF-1 and IGF binding proteins 1, 2, and 3 were measured by ELISA’s.

Results

Exercise adherence was 88% and the dropout rates for the exercise and control groups were 21.7% and 14.5%, respectively. There was a small significant change from baseline in IGFBP-3 concentrations. IGFBP-3 levels decreased in controls and increased in exercisers. The between group difference was significant. No other changes were noted.

Conclusion

Sixteen weeks of exercise had minimum or no effect on IGF proteins of young women.

Impact

Our study supports findings from previous studies conducted in older populations and raises the question of what type of intervention is needed to change circulating levels of IGF proteins in humans.

Introduction

Insulin-like growth factor 1 (IGF-1) and IGF binding protein 3 (IGFBP-3) have been associated with increased risk of premenopausal breast cancer (1). Recently, results from a pooled data analysis of 17 prospective studies revealed that IGF-1 levels were positively associated with estrogen receptor positive breast cancer, regardless of menopausal status (2). It is unclear whether exercise training affects circulating levels of IGF-1 and binding proteins. Previous exercise trials in postmenopausal women have reported mixed results (3–5). In young women, although strength training has been found to decrease IGF-1 and insulin levels (6), there is a lack of large longitudinal studies examining the effects of aerobic exercise on IGF proteins. We conducted a randomized clinical trial to examine the effects of exercise training on insulin, glucose, and IGF axis proteins in sedentary young women. We hypothesized that exercise training would lower IGF-1, insulin, glucose and the molar ratio of IGF-1/IGFBP-3 and increase binding protein levels in the direction associated with decreased breast cancer risk.

Methods

The Women in Steady Exerciser Research (WISER) study was a randomized clinical trial comparing the effects of a 16-week aerobic exercise intervention versus no exercise on markers associated with breast cancer risk (7). The study was approved by the University of Minnesota Human Subjects Review Committee. Participants were healthy women aged 18–30 years, with a BMI between 18.5–40 kg/m², sedentary, not taking hormonal contraceptives, and not dieting. Women were randomly assigned to the exercise intervention (n = 166) or the control group (n = 153).

The exercise intervention consisted of 30 minutes of weight-bearing aerobic exercise, five times per week, at a specified intensity based on age-predicted heart rate maximum (HRmax). Exercise intensity increased every four weeks to reach 80–85% of HRmax. The control group was asked to maintain their usual levels of physical activity.

At baseline and 16-weeks the following were collected: demographic information, anthropometrics, body composition via DXA (Lunar Radiation Corp., Madison, WI), dietary intake via 3-day food records, health habits, fitness level via sub-maximal tests, and blood samples, which were collected in the morning during the follicular phase of participants’ cycles.

Fasting glucose and insulin levels were measured at the Fairview University Diagnostic Laboratories (Minneapolis, MN). IGF proteins were measured by ELISAs (DSL, Webster, TX). Intra- and inter-assay coefficients of variation for IGF-1, IGFBP-1, IGFPB-2, and IGFBP-3 were 1.8%, 1.9%, 2.2%, and 1.3% and 7.2%, 9.3%, 10.6%, and 4.9%, respectively. All samples from the same participant were analyzed in the same batch.

Unadjusted comparisons of demographic characteristics were performed by t-tests and chi-squares for continuous and categorical variables, respectively. Comparisons between groups were adjusted for age and BMI strata with a general linear model. This was an intent-to-treat analysis. Insulin and IGFBP-1 were analyzed on the log scale; changes in outcome variables were analyzed on the original scale. We had 80% power to detect a difference in change from baseline to 16 weeks between exercisers and controls of 20.7 ng/L for IGF-1, 1.29 mU/L for insulin, 0.146 mmols/L for glucose and 4.97, 51.4, and 161 ng/mL for IGFBP-1, -2, and -3, respectively. All analyses were performed with SAS software (version 9.2, SAS Institute, Cary, NC).

Results

We randomized 391 participants into the study and 319 successfully completed the study. The dropout rates for the exercise and control groups were 21.7% and 14.5%, respectively. Three participants (two controls and one exerciser) were excluded from analysis because of unusually high follow-up insulin levels (i.e., more than a 5-fold increase from baseline).

There were no baseline differences in demographic characteristics and other variables between groups (Table 1). Fitness increased by 0.91 METs in the exercisers compared to 0.14 METs in controls (p < 0.0001). Exercisers completed an average 92% of assigned minutes of exercise, and 140 women (86%) exercised for at least 14 weeks.

Table 1.

Baseline characteristics of controls and exercisers^*.

	Controls (n = 151)	Exercisers (n = 165)	p
Age (years )	25.3 ± 3.5	25.4 ± 3.4	0.8
Body Fat (%)	36.2 ± 8.3	36.3 ± 8.7	0.86
Weight (kg)	67.7 ± 14.7	67.4 ± 14.6	0.86
Fitness (METS)	7.0 ± 1.5	6.9 ± 1.5	0.59
Ethnicity
Hispanic	6 (4.0)	8 (4.8)	0.7
Non-hispanic	145 (96.0)	157 (95.2)
Race
American Indian	0	1 (0.6)	0.59
Asian/Pacific Islander	26 (17.2)	20 (12.1)
Black	11 (7.3)	13 (7.9)
White	106 (70.2)	124 (75.1)
Other/multi-race	8 (5.3)	7 (4.2)
Education
Less than college	49 (32.5)	53 (32.0)	0.27
College degree	63 (41.7)	56 (34.0)
Graduate/professional degree	39 (25.8)	56 (34.0)
Marital Status
Never married	122 (80.8)	137 (83.0)	0.62
Married/partnered	26 (17.2)	25 (15.2)
Divorced/separated	3 (2.0)	3 (1.8)
BMI
≤ 25 kg/m2	99 (65.5)	101 (61.2)	0.2
25–30 kg/m2	30 (19.9)	46 (27.9)
≥ 30 kg/m2	22 (14.6)	18 (10.9)

^*Values are presented as mean ± SD for continuous variables or n (%) for categorical variables.

IGFBP-3 levels decreased in controls and increased in exercisers with the difference in change between groups being significant (Table 2). There were no differences between groups in the other outcomes.

Table 2.

Changes in glucose, insulin, insulin resistance and IGF proteins for controls versus exercisers^a.

	Baseline	4 months	Change at 4 months	% Change at 4 months
Glucose (mmols/L)
Control (n = 151)	4.44 ± 0.03	4.40 ± 0.03	−0.76 ± 0.7	−0.50 ± 0.9
Exercise (n = 165)	4.52 ± 0.03	4.50 ± 0.03	−0.39 ± 0.7	0.08 ± 0.8
p-value	0.12	0.04	0.696	0.63
Insulin (mU/L)
Control (n = 151)	6.22 (5.7 – 6.8)	6.31 (5.8 – 6.9)	0.10 ± 0.3	19.0 ± 5.6*
Exercise (n = 165)	6.14 (5.6 – 6.7)	5.82 (5.3 – 6.3)	−0.51 ± 0.3	8.2 ± 5.4
p-value	0.835	0.194	0.186	0.158
HOMA index
Control (n = 151)	1.49 ± 0.08	1.49 ± 0.07	0.00 ± 0.07	20.6 ± 6.2
Exercise (n = 165)	1.48 ± 0.07	1.37 ± 0.06	−0.12 ± 0.07	9.5 ± 5.9
p-value	0.954	0.229	0.236	0.186
QUICKI
Control (n = 151)	0.393 ± 0.0	0.375 ± 0.0	−0.02 ± 0.0*	−4.50 ± 0.8*
Exercise (n = 165)	0.391 ± 0.0	0.378 ± 0.0	−0.01 ± 0.0*	−3.30 ± 0.7*
p-value	0.158	0.455	0.247	0.254
IGF-1 (ng/mL)
Control (n = 151)	394.7 ± 7.8	382.1 ± 7.8	−11.1 ± 5.5	−1.63 ± 1.5
Exercise (n = 165)	386.0 ± 7.4	377.3 ± 7.3	−9.8 ± 5.2	−0.84 ± 1.4
p-value	0.417	0.652	0.867	0.695
IGFBP-1 (ng/mL)
Control (n = 151)	21.1 (18.9 – 23.6)	20.5 (18.1 – 23.1)	−0.25 ± 1.3	15.4 ± 7.1*
Exercise (n = 165)	24.5 (22.0 – 27.2)	22.9 (20.4 – 25.6)	−0.90 ± 1.3	12.2 ± 6.8
p-value	0.059	0.181	0.712	0.742
IGFBP-2 (ng/mL)
Control (n = 151)	347.6 ± 20.6	408.0 ± 21.5	52.2 ± 14.8	37.4 ± 24.5*
Exercise (n = 165)	370.6 ± 19.4	414.3 ± 20.3	37.4 ± 14.4	51.8 ± 23.8*
p-value	0.413	0.828	0.412	0.638
IGFBP-3 (ng/mL)
Control (n = 151)	4942.1 ± 69.4	4837.1 ± 68.7	−103.1 ± 42.6	−1.49 ± 0.9
Exercise (n = 165)	4863.6 ± 65.5	4883.9 ± 64.8	18.7 ± 40.7	0.99 ± 0.9
p-value	0.407	0.617	0.035	0.051
IGF-1:IGFBP-3†
Control (n = 151)	0.29 ± 0.00	0.28 ± 0.00	−0.317 ± 0.3	−0.45 ± 1.1
Exercise (n = 165)	0.29 ± 0.00	0.28 ± 0.00	−0.817 ± 0.3	−1.85 ± 1.0
p-value	0.959	0.543	0.22	0.334

^aValues are expressed either as mean ±SE or geometric mean (lower CI – upper CI) for baseline data and as mean ±SE for changes.

^*Significant change from baseline within treatment group at p<.05.

^†Values for change at 4 months were multiplied by 100 to facilitate presentation.

Discussion

This randomized trial suggests that 16 weeks of aerobic exercise, 30 minutes/day, 5 days/week, has minimal or no effect on levels of insulin, glucose, IGF-1 and IGF binding proteins among young, healthy eumenorrheic women. The WISER study strengths include the type of exercise intervention chosen, which was based on current public health guidelines, and the excellent retention and adherence rates. Furthermore, previous studies examining the effects of exercise on IGF axis proteins have focused on postmenopausal women and found mixed results (3–4). The WISER study is the first large trial to assess the effects of aerobic exercise on IGF axis proteins in young women. Though we found minimal effects, it may be that longer exercise interventions or alternative approaches to the current physical activity interventions are needed to change the levels of these markers in young women.