Basal growth hormone concentration increased following a weight loss focused dietary intervention in older overweight and obese women

Abstract

Objectives

Growth and sex steroid hormones decrease with aging and obesity. The effect of dietary weight loss and exercise training lifestyle interventions was examined on hormones as well as determining their relationships with physical function in older obese and overweight adults.

Design

Individuals were randomized into one of four 18 month interventions: Healthy Lifestyle (HL), Exercise, Diet, and Exercise-Diet.

Setting

Clinical research setting with facility based exercise and nutrition education and behavior classrooms.

Participants

Older (≥60 yrs) overweight and obese (BMI≥28 kg/m²) adults with knee osteoarthritis (n=309) were recruited for the study.

Intervention

Weight loss goal for Diet groups was ≥5%. Exercise groups trained (mostly walking and resistance training) 3 days/week for 60 min/session.

Measurements

Body weight, growth hormone (GH), corticosterone, sex-hormone binding globulin (SHBG), testosterone, and dehydroepiandrosterone (DHEA) were measured at baseline, 6, and 18 months. Physical function was determined through performance task (6- min walking distance) and self-reported questionnaires (Western Ontario McMaster University Osteoarthritis Index-WOMAC) at similar time points.

Results

Diet, Exercise, and Exercise-Diet groups lost 4.9%, 3.5%, and 6.2% of their weight at 18 months, respectively. There was a significant diet treatment effect on GH levels in women as higher concentrations of this hormone were apparent following dietary weight loss intervention (p=0.01). No other hormones were affected by either diet or exercise treatments in men or women. A significant inverse correlation between baseline 6-minute walking distance and SHBG (r=−0.33) was found in men.

Conclusion

The increase in basal GH levels from the diet treatment in women suggests that this lifestyle behavior intervention may mitigate the age- and obesity-related decreases in growth hormone levels, to help preserve muscle mass, strength, and physical function in older adults.

Introduction

Altered blood concentrations of sex and anabolic hormones are associated with many aging-related chronic diseases, including among others, reductions in skeletal muscle mass and strength (1). Reduced anabolic stimulus from reduced levels of growth hormone (GH) and testosterone may lead to a reduction in skeletal and muscle mass, and an increase in abdominal visceral fat (2). As a consequence, sarcopenic obesity may ensue, resulting in impairments in physical function, mobility disability, and exercise capacity (3). This is consistent with findings that individuals with GH deficiency have impairments in physical function and increased disability (4). Evidence suggests that normal aging lowers blood concentrations of growth and sex steroid hormones and binding proteins (5., 6., 7., 8.). Obesity, in addition to aging, has profound effects on the metabolism and concentrations of hormones. For example, testosterone and DHEA concentrations are lower in obese vs. nonobese men with significant inverse relationships between hormones and visceral fat levels (9). Finally, impairments in physical function and performance (10) are more prevalent in older adults, consistent with reports that one-third of older adults report having some level of disability or difficulties in physical function (11).

Lifestyle modifications to increase physical activity and diet restriction to lower body weight are frequently the first line of therapy to reduce risk of chronic diseases. However, the effect of these behavior changes on circulating concentrations of sex and anabolic hormones is not definitely known, especially in older individuals that do not have frank hormone deficiencies. cross-sectional and longitudinal trials in younger cohorts indicate that weight loss and exercise training improve hormonal profiles. Rapid and sustained weight loss in younger obese individuals increase levels of total testosterone, free testosterone, sex-hormone binding globulin (sHBG) and GH (12, 13). Growth hormone concentrations increase following resistance and endurance exercise in women and men, although this effect was found to be diminished in older men (14, 15).

There have only been a very limited number of studies that have examined the effect of weight loss and exercise programs in older obese adults in a randomized clinical trial design. These have been limited in that they are only for a short duration or only in those without frank hormonal deficiency. Therefore, the primary aim of this study was to examine the effects of randomization to an 18-month intervention of dietary focused weight loss, exercise training interventions, their combination, or a control group on hormone levels in older overweight and obese men and women. cross-sectional analysis at baseline was also performed to further support the relationship between hormone concentrations with body mass index (BMI) and physical function.

Methods

Study Population

Study participants were enrolled in the Arthritis, Diet, and Activity Promotion Trial (ADAPT). complete details of this study design and primary outcome measures are published elsewhere (16, 17). iRB approval was acquired and informed consent was obtained from participants. ADAPT compared the effects of an 18 month dietary weight-loss intervention (Diet), an exercise intervention (Exercise), a combined exercise- dietary weight loss intervention (exercise-Diet) and a healthy lifestyle control group (HL), on self-reported disability and physical function in older, overweight and obese, sedentary adults with knee osteoarthritis (OA). community dwelling older adults (n=316; > 60 years of age) with symptomatic knee OA were recruited. Blood was obtained from 309 of the participants for this analysis. Additional inclusion criteria included BMI > 28.0 kg-m-2, sedentary lifestyle, and self- reported difficulty in performing at least one of the following activities attributed to knee pain: lift and carrying groceries, walking one-quarter mile, getting in and out of a chair, or going up and down stairs.

Interventions

Both Diet and Exercise-Diet groups were prescribed similar dietary weight loss intervention strategies. The weight loss goal for these two groups was a mean loss > 5% of initial body weight. Details of the intervention are described elsewhere (16). Participants were individually counseled on reducing energy intake by ~250-500 calories per day to achieve the desired weight loss. They met in group and individual sessions throughout the study in a 3:1 ratio. Primary strategies included cognitive behavioral based treatment with emphasis on reducing portion sizes and fat intake.

The exercise program was conducted three days per week for 60 minutes per session and was similar for the Exercise and Exercise-Diet groups. The exercise program consisted of a warm-up phase (5 minutes), an aerobic phase (15 minutes), a strength phase (20 minutes), a second aerobic phase (15 minutes), and a cool-down phase (5 minutes). The primary mode of aerobic training consisted of walking. The exercise intensity for the aerobic exercise was 50-85% of the heart rate reserve using the symptom limited maximum heart rate obtained from a graded exercise test (GXT). Lower body strength training was also employed.

Participants randomized to the HL group met monthly for 1 hour during the first 3 months of the trial. Topics for these sessions included osteoarthritis, obesity, and exercise. In addition, phone contacts were performed on a monthly (for months 4-6) and bimonthly (months 7-18) basis.

Measurements

Descriptions and time course of all measurements obtained in ADAPT are reported by Miller and coworkers(16). At baseline, 6 months and 18 months testing visits, participants reported in the morning after a 12-hour fast to the General Clinical Research Center (GCRC) of Wake Forest University Medical center for blood draws via the antecubital vein and body weight and height.

Serum was used for the measurement of the hormones and binding protein (corticosterone, GH, DHEA, testosterone, and sHBG). All hormones were measured using an automated immunoanalyzer (IMMULITE®, Diagnostics Products corporation, Los Angeles, cA). sensitivities and ranges, respectively, for the hormones and binding protein, were corticosterone, 5.52 nmol/L and 27.6-1380.0; GH 0.01 µg/L and 0-40; testosterone 0.35 nmol/L and 0.7-55.5; DHEA 69.4 nmol/L and 104.1-3470.0; and SHBG 0.02 nmol/L and 0-180 nmol/L. All samples were measured in duplicate and the average of the two values was used for data analyses.

Self-reported physical function, pain, and stiffness were measured using the 24 item Western Ontario and McMaster University Osteoarthritis Index (WOMAC) (18). The WOMAc has been validated, and is recommended by the Osteoarthritis Research Society as the measure of choice (18). Participants respond to specific questions rating their pain, stiffness, and difficulty to perform specific tasks on a 1 (none) to 5 (extreme). Physical performance was determined using the 6-minute walk distance. Participants were instructed to walk as far as possible in a 6-minute time period on an established course. They were not allowed to carry a watch and were not provided with feedback during the trial.

Statistical Analysis

Simple statistics such as means, standard errors, and percent distributions were computed on baseline descriptive measures, overall and by study group. Means and standard errors were calculated by gender and group for baseline hormone measures. Group differences at baseline were assessed using an ANOVA analysis for continuous measures and chi-square testing for categorical measures. A repeated measures analysis was performed using mixed modeling to look at the differences between the study groups and the change in hormones over the length of the study. Additional covariates in the models included race, age, and baseline measures of hormones and BMI. since there were differences between males and females on some outcome measures, analyses were done stratified on gender. A second analysis for examining main effects of diet and exercise interventions on hormone levels was performed. Mixed model based least square means and standard errors at each visit were computed by gender and group. Log- transformations were used on hormone outcomes presenting skewed distributions. Lastly, we looked at baseline relationships by gender between hormones and BMI and function measures using spearman correlation. significance was determined using a 0.05 level. All analyses were conducted using SAS version 9.1 (Cary, NC).

Results

Baseline characteristics across each of the four arms of the study are presented in Table 1. There were no significant differences among the groups at baseline for age, gender, race, body weight, or BMI. The mean age for participants at the start of the study was 68.6 (0.4) years with 72% women and 76% white. Baseline BMi was 34.3 (0.3) kg/m² with an average body weight of 93.9 (1.0) kg across all groups.

Table 1.

Baseline characteristics by groups

Mean (SD)	Overall	Healthy Lifestyle	Diet	Exercise	Diet - Exercise	P value
Number	N = 309	N = 76	N = 80	N = 79	N = 74
Age (yrs)	68.6 (0.4)	68.7 (0.7)	67.9 (0.6)	69.0 (0.7)	68.7 (0.8)	0.62
Gender (% females)	72.2	67.1	71.3	76.0	74.3	0.63
Race (%)
White	75.7	79.0	71.3	74.7	78.4	0.85
Black	22.3	19.8	27.5	22.8	18.9
Other	2.0	1.2	1.2	2.5	2.7
Body Weight	93.9 (1.0)	95.8 (2.2)	95.2 (1.7)	92.1 (1.7)	92.2 (2.1)	0.39
(kg)
Body Mass	34.3 (0.3)	34.3 (0.6)	34.5 (0.6)	34.2 (0.6)	34.2 (0.7)	0.98
Index (kg-m−2)

values are presented as Means ± SEM unless indicated; N’s represent the maximum number of participants within each grouping

At 6 months, the exercise-Diet group had lost 6.3% (3.1) of their initial body weight, and maintained that level at 18 months (6.2% (3.1)). The Diet only group also showed significant weight loss at 6 months (3.9% (2.6)) and at 18 months (4.9%, (2.7)). Although the Exercise group did not show weight loss at 6 months (0.7% (2.8)), by 18 months they had lost 3.5% (2.7) of their initial weight. The Healthy Lifestyle controls also lost 1.1% (3.4) and 1.6% (3.4) at 6 and 18 months, respectively.

Baseline levels of GH, testosterone, and SHBG were significantly different among men and women; testosterone was higher in men, and GH and sHBG were higher in women. Thus, hormone results are presented and analyzed separately based on gender. at baseline there were no differences among treatment groups in circulating levels of hormones or binding proteins for men or women (P values not shown for baseline comparisons). comparisons across time and interventions showed no differences in hormones or binding protein levels between treatment groups or data collection periods in men. However, in women, there were 2 notable significant effects. there was an increase in cortisol concentrations at 18 months compared to 6 months (p=0.01) (Table 2). Additionally, and the most marked finding from this study, was a treatment effect on GH (Table 2) (p=0.01). specifically, both Diet related treatment groups demonstrated higher levels of GH compared to the non-Diet treated groups (Exercise and HL) (all p<0.05). there was no other treatment effect on hormones for diet or exercise interventions.

Table 2.

serum hormone and binding protein levels at baseline, 6- months and 18-months by gender for each intervention group

	Healthy Lifestyle	Diet	Exercise	Diet - Exercise	P-value for Comparing Groups Across Follow-Up	P-value for Comparing Across Time
Men (N)	25	23	19	19
Cortisol (nmol/L)
Baseline	289.7±24.8	278.7±30.3	248.3±22.1	253.8±19.3	.88	.37
6-Months	229.0±22.1	267.6±22.1	240.0±24.8	251.1±27.6
18-Months	278.7±22.1	220.7±22.1	278.7±24.8	273.1±27.6
DHEA (nmol/L)
Baseline	279.0±34.0	263.4±37.5	250.5±28.1	266.8±27.1	.64	.26
6-Months	289.1±29.8	231.8±30.2	228.3±30.2	245.3±35.7
18-Months	282.8±30.5	255.0±28.8	284.2±33.0	251.2±34.4
Growth Hormone µg/L
Baseline	3.4±1.2	2.4±0.9	2.4±1.0	6.9±3.0	.57	.33
6-Months	3.5±2.8	2.6±3.0	7.4±2.5	11.1±3.3
18-Months	5.8±2.8	3.4±3.5	3.3±2.7	3.0±4.0
Testosterone (nmol/L)
Baseline	9.9±0.8	9.4±1.0	9.5±1.0	10.6±1.1	.77	.76
6-Months	9.2±0.8	10.4±0.8	10.1±0.9	10.2±0.9
18-Months	9.7±0.8	9.6±0.8	9.3±0.9	10.7±1.0
SHBG nmol/L
Baseline	35.7 ± 2.3	35.0 ± 3.1	35.2 ± 3.6	39.1 ± 3.6	.89	.69
6-Months	35.9 ± 4.2	39.6 ± 4.1	38.9 ± 4.4	40.0 ± 4.9
18-Months	42.3 ± 3.9	39.4 ± 3.7	35.1 ± 4.6	41.5 ± 4.7
Women (N)	51	57	60	55
Cortisol (nmol/L)
Baseline	264.9±13.8	278.7±16.6	251.1±13.8	289.7±13.8	.60	.01
6-Months	231.8±13.8	242.8±13.8	237.3±13.8	245.6±16.6
18-Months	251.1±16.6	286.9±16.6	251.1±13.8	262.1±16.6
DHEA (nmol/L)
Baseline	232.5±25.7	231.8±19.4	210.3±20.1	206.1±17.4	.66	.60
6-Months	214.8±15.6	217.2±14.6	229.0±20.5	222.1±18.7
18-Months	197.8±17.4	242.6±16.7	228.3±17.7	231.4±20.1
Growth Hormone (µg/mL)
Baseline	10.3±2.5	9.5±2.4	9.0±1.8	7.7±1.1	.01	.11
6-Months	7.7±2.4	15.3±2.3	3.0±2.5	12.8±2.5
18-Months	5±2.7	2.7±10.3	4.8±2.6	11.9±2.7
Testosterone (nmol/L)
Baseline	1.7±0.2	1.8±0.2	1.7±0.2	1.7±0.1	.26	.40
6-Months	1.7±0.3	1.8±0.2	1.3±0.2	1.7±0.3
18-Months	1.8±0.3	2.0±0.2	1.4±0.3	1.8±0.3
SHBG (nmol/L)
Baseline	60.4 ± 6.2	64.2 ± 5.4	57.6± 5.1	76.8 ± 6.7	.70	.55
6-Months	66.7 ± 3.9	70.5 ± 3.8	65.2 ± 4.1	74.1 ± 4.8
18-Months	69.6 ± 4.3	71.1 ± 4.3	65.4 ± 4.2	61.8 ± 4.8

DHEA, dehydroepiandosterone; sHBG, sex hormone binding globulin; values are presented as Means ± SEM.

Physical function measures at baseline, separated by gender are shown in Table 3. This is for the 6 minute walk distance as well as the self-reported WOMAC questionnaire. Furthermore, correlations were performed between baseline measures of function and hormones for men and women (Tables 4). A significant negative correlation was apparent between plasma levels of sHBG and distance walked (r=-0.33) in men, such that those with higher levels of sHBG walked less distance. No other significant associations between hormones and function were observed, although there was a trend for an association between testosterone and WOMAC stiffness in men (r=-0.21; p=0.08) and testosterone and WOMAc function in women (r=-0.18; p=0.09).

Table 3.

Physical function measures at baseline

	Men	Women
Distance Walked (m)	1516.14 ± 24.53	1317.57 ± 17.44
WOMAC Function (units)	1.27 ± 0.07	1.50 ± 0.05
WOMAC Pain (units)	1.26 ± 0.07	1.45 ± 0.05
WOMAC Stiffness (units)	1.66 ± 0.08	1.94 ± 0.05

values are means ± SEM

Table 4.

Associations of functional measures with hormones and binding protein at baseline for men (A) and women (B)

A. Men	Distance Walked	WOMAC Function	WOMAC Stiffness	WOMAC Pain
Cortisol	r = 0.09	−0.01	−0.09	0.07
	p = 0.45	0.95	0.45	0.54
	n = 70	72	72	72
DHEA	−0.07	−0.14	−0.14	−0.08
	0.63	0.30	0.28	0.55
	57	59	59	59
Growth Hormi	ne -0.03	−0.16	−0.03	−0.18
	0.85	0.26	0.82	0.21
	50	51	51	51
SHBG	−0.33	0.06	0.01	0.16
	0.01	0.61	0.92	0.17
	70	72	72	72
testosterone	0.04	−0.18	−0.21	0.04
	0.74	0.13	0.08	0.73
	70	72	72	72
B.	Distance	WOMAC	WOMAC	WOMAC
Women	Walked	Function	Stiffness	Pain
Cortisol	r = -0.11	0.01	0.06	0.03
	p = 0.14	0.85	0.46	0.68
	n=168	188	188	188
DHEA	0.13	0.09	0.11	−0.04
	0.22	0.37	0.28	0.68
	89	102	102	102
Growth Horm	ne -0.13	−0.02	0.003	−0.02
	0.11	0.78	0.97	0.80
	151	169	169	169
SHBG	−0.05	−0.11	0.01	−0.08
	0.51	0.16	0.93	0.31
	158	177	177	177
testosterone	−0.08	−0.18	−0.05	−0.15
	0.45	0.09	0.65	0.14
	81	92	92	92

Discussion

This study examined the effects of lifestyle behavior based interventions incorporating dietary induced weight loss and exercise training on growth and sex hormones and binding protein levels in older overweight and obese men and women. In addition, cross-sectional analysis was performed to assess the relationships between hormone concentrations and physical function measures at baseline. Previous research has shown that aging, weight status, and physical activity alter hormone concentrations, which may contribute to diabetes, skeletal muscle loss, and visceral fat accumulation, among other conditions(19., 20., 21.). Since initial treatment of these comorbidities involves lifestyle modifications of diet and exercise, it was of interest to determine if interventions that target dietary and physical activity behaviors improve hormones and binding protein profiles. Our findings are noteworthy in that they provide randomized controlled trial evidence that the main effect of an 18-month dietary induced weight loss intervention increased systemic concentrations of GH in older overweight and obese women compared to women randomized to the non- Diet intervention groups. The potential impact of these findings is to lessen the reductions or even improve changes associated with aging, obesity, and reductions in hormones, that lead to decreases in muscle mass and strength and detriments in physical function. Whereas some clinicians continue to be hesitant in prescribing weight loss for their obese older adults, these data provide further evidence for improvements in health indices accompanying weight loss in this population.

Previous evidence demonstrates a robust relationship between decreases in GH and changes in body composition and physical function that occur with aging(1, 22). However, attempts to restore GH levels through exogenous administration of the peptide to older adults are presently debated due to equivocal findings (23). Supplementation of GH in pathologic deficiency states has been shown to be effective in certain animal and human models in restoring muscle mass, with greater effects occurring when incorporating exercise with the supplement (24, 25). However, in non-GH deficient models, the findings are controversial with some showing improvements in protein synthesis (26) and muscle mass (27, 28), whereas others showed no effect of supplementation on muscle mass, strength, or performance (29). Additionally, supplementation of GH produces significant side effects, including edema, gynecomastia, disturbances in glucose homeostasis, and joint pain (30, 31). Because of these untoward actions, demonstrating increases in GH levels through behavioral interventions is significant. In earlier work, we showed that both exercise and diet interventions in this cohort produced improvements in physical function in men and women combined; however, no measures of strength or body composition were performed in this study (17). Thus, probable mechanisms for physical function changes in women in the diet intervention groups may be related to the GH improvements. This is an area of future research.

The underlying reason for the gender effect from the diet treatment with GH is not known. In nonpathological states, women and men show decreases in circulating daytime and nighttime concentrations of GH with age (32, 33); growth hormone secretion is pulsatile with maximal secretion occurring in the night. However, Latta et al., showed no differences in 24-hour GH secretion in healthy older men and women, but did show that men had greater nighttime secretion of GH compared to women (34). Only basal GH concentrations were obtained in our study, which limits the interpretation of the findings and may explain our lack of a finding of intervention changes in GH in males. Determining 24 hour sampling would provide a more complete picture of the effect of the intervention on the hormone, including capturing peak levels; the largest secretion occurs at the onset of the sleep and represents the majority of the daily output. An additional limitation for this study is the lack of measurement of other components of the GH axis, including insulin-like growth factor 1 (IGF-1). Recent work has shown that IGF-1 is lower in women vs. men, decreases with age and higher levels of BMi, and is lower in individuals with increasing number of metabolic syndrome factors (35). Furthermore, higher IGF-1 levels have been associated with better physical function measures in older obese adults (36). in light of this earlier work and our current findings, this is a potentially critical area of research to pursue.

The lack of change from the interventions for SHBG and testosterone is in contrast to an earlier cross-sectional study which showed exercise trained older men had higher sHBG and testosterone levels than age-matched sedentary men (37). selected studies have investigated hormone changes following exercise training programs, although they are limited by being cross-sectional in design, their study length, type of exercise performed, or age of cohort. Results from a heavy resistance training program showed significant improvements in strength in older men and women after 6 months of training (38), but similar to our findings, no changes in the hormone profile which included testosterone, DHEA, GH, cortisol, or SHBG, were found from the intervention.

Obesity has been shown to reduce steroid hormones and binding proteins; reductions in testosterone and SHBG were shown in men who became obese during a follow-up aging study (39). in support of this earlier work, others found that both testosterone and SHBG were negatively correlated with measures of body fat (6). The impact that weight loss may have on these hormone levels is limited and the few published studies have focused on a younger cohort with mean ages of 30-45 years (12) compared to over 65 years in the current analysis. For example, testosterone and SHBG increased with rapid and sustained weight loss in men with metabolic syndrome abdominal obesity and placed on a very low calorie diet for 9 weeks followed by a 12 month maintenance period (12).

Because obesity is a risk factor for osteoarthritis of the knee (40), and this joint disease is the leading cause of functional impairment (41), it was of interest to examine the relationships of the anabolic and sex hormones with physical function measures in this obese and overweight older adult cohort. We hypothesized that individuals with lower levels at baseline of hormones would have worse function. This was supported from results showing that SHBG is inversely related to distance walked (r=-0.33), indicating that higher levels of SHBG are associated with a shorter walking distance. SHBG binds to estrogen and testosterone, thereby reducing the active, free form of the hormone. This was the only significant relationship observed between the hormones and functional measures at baseline. Previous work by O’Donnell and colleagues showed that in older men (55-85 years), low levels of testosterone and DHEA were related to poor physical performance(42). They further suggested that increasing hormone levels above a threshold does not confer additional benefit. this may explain our lack of associations as the current concentrations of the hormones at baseline may have been higher than this threshold, thereby eliminating significant associations with functional measures.

In summary, aging and obesity have been shown to affect levels of circulating growth and sex anabolic hormones. Our results demonstrate an 18 month dietary-induced weight loss intervention increases basal levels of growth hormone in women. the clinical impact of this is not known as there is lack of a consensus on the impact of raising baseline concentrations of this hormone in older adults, although it would suggest a potential improvement, or a lessening in detriments in muscle mass and strength present with aging. Furthermore, exercise training had no impact on hormone levels in men or women. The lack of association between function measures and hormones at baseline may be a reflection of hormones being above a threshold.

Funding and Support

This project was funded by the Claude D. Pepper Older Americans (Grant 5P60AG10484-00) and the General Clinical Research Center (Grant M01-RR07122).

Disclosure Statement

The authors have nothing to disclose.