Effects of individual and combined dietary weight loss and exercise interventions in postmenopausal women on adiponectin and leptin levels

Authors

  • C. Abbenhardt,

    1. Division of Preventive Oncology, National Center for Tumor Diseases and German Cancer Research Center, Heidelberg, Germany
    2. Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA, USA
    Search for more papers by this author
  • A. McTiernan,

    1. Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA, USA
    2. School of Medicine, University of Washington, Seattle, WA, USA
    3. School of Public Health, University of Washington, Seattle, WA, USA
    Search for more papers by this author
    • These authors are dual senior authors.
  • C. M. Alfano,

    1. Office of Cancer Survivorship, National Cancer Institute/National Institutes of Health, Bethesda, MD, USA
    Search for more papers by this author
  • M. H. Wener,

    1. Department of Laboratory Medicine, University of Washington, Seattle, WA, USA
    Search for more papers by this author
  • K. L. Campbell,

    1. Faculty of Medicine, University of British Columbia, Vancouver, BC, Canada
    Search for more papers by this author
  • C. Duggan,

    1. Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA, USA
    Search for more papers by this author
  • K. E. Foster-Schubert,

    1. School of Medicine, University of Washington, Seattle, WA, USA
    Search for more papers by this author
  • A. Kong,

    1. Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA, USA
    2. Cancer Education and Career Development Program, University of Illinois at Chicago, Chicago, IL, USA
    Search for more papers by this author
  • A. T. Toriola,

    1. Division of Preventive Oncology, National Center for Tumor Diseases and German Cancer Research Center, Heidelberg, Germany
    2. Department of Surgery, Division of Public Health Sciences and Siteman Cancer Center, Washington University School of Medicine, St. Louis, MO, USA
    Search for more papers by this author
  • J. D. Potter,

    1. Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA, USA
    2. School of Public Health, University of Washington, Seattle, WA, USA
    3. Centre for Public Health Research, Massey University, Wellington, New Zealand
    Search for more papers by this author
  • C. Mason,

    1. Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA, USA
    Search for more papers by this author
  • L. Xiao,

    1. Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA, USA
    Search for more papers by this author
  • G. L. Blackburn,

    1. Division of Nutrition, Harvard Medical School, Boston, MA, USA
    Search for more papers by this author
  • C. Bain,

    1. Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA, USA
    Search for more papers by this author
  • C. M. Ulrich

    Corresponding author
    1. Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA, USA
    2. School of Public Health, University of Washington, Seattle, WA, USA
    • Division of Preventive Oncology, National Center for Tumor Diseases and German Cancer Research Center, Heidelberg, Germany
    Search for more papers by this author
    • These authors are dual senior authors.

Errata

This article is corrected by:

  1. Errata: Corrigendum Volume 276, Issue 4, 418, Article first published online: 20 September 2014

Correspondence: Cornelia M. Ulrich PhD, Division of Preventive Oncology Preventive Oncology, National Center for Tumor Diseases/German Cancer Research Center, Im Neuenheimer Feld 460, 69120 Heidelberg, Germany.

(fax: +49 6221 56 5231; e-mail: neli.ulrich@nct-heidelberg.de).

Abstract

Background

Excess body weight and a sedentary lifestyle are associated with the development of several diseases, including cardiovascular disease, diabetes and cancer in women. One proposed mechanism linking obesity to chronic diseases is an alteration in adipose-derived adiponectin and leptin levels. We investigated the effects of 12-month reduced calorie, weight loss and exercise interventions on adiponectin and leptin concentrations.

Methods

Overweight/obese postmenopausal women (n = 439) were randomized as follows: (i) a reduced calorie, weight-loss diet (diet; N = 118), (ii) moderate-to-vigorous intensity aerobic exercise (exercise; N = 117), (iii) a combination of a reduced calorie, weight-loss diet and moderate-to-vigorous intensity aerobic exercise (diet + exercise; N = 117), and (iv) control (N = 87). The reduced calorie diet had a 10% weight-loss goal. The exercise intervention consisted of 45 min of moderate-to-vigorous aerobic activity 5 days per week. Adiponectin and leptin levels were measured at baseline and after 12 months of intervention using a radioimmunoassay.

Results

Adiponectin increased by 9.5% in the diet group and 6.6% in the diet + exercise group (both ≤ 0.0001 vs. control). Compared with controls, leptin decreased with all interventions (diet + exercise, −40.1%, P < 0.0001; diet, −27.1%, P < 0.0001; exercise, −12.7%, = 0.005). The results were not influenced by the baseline body mass index (BMI). The degree of weight loss was inversely associated with concentrations of adiponectin (diet, P-trend = 0.0002; diet + exercise, P-trend = 0.0005) and directly associated with leptin (diet, P-trend < 0.0001; diet + exercise, P-trend < 0.0001).

Conclusion

Weight loss through diet or diet + exercise increased adiponectin concentrations. Leptin concentrations decreased in all of the intervention groups, but the greatest reduction occurred with diet + exercise. Weight loss and exercise exerted some beneficial effects on chronic diseases via effects on adiponectin and leptin.

Background

Being overweight or obese, and having a sedentary lifestyle may account for as much as 80% of the most common chronic diseases, including cardiovascular disease, hypertension, cancer and type 2 diabetes [1]. Obesity is a risk factor for cardiovascular disease and is associated with an increased risk of morbidity and reduced life expectancy [2]. In addition, obesity drives diabetic and vascular mortality [3, 4]. Higher levels of regular, moderate-intensity physical activity are associated with a reduced risk of several cancers, including cancers of the breast, colon and endometrium [5, 6]. Prospective studies suggest that women who lose weight and maintain the weight loss experience a reduction in breast cancer risk [6-9]. Although the biologic mechanisms that mediate the association of obesity, sedentary lifestyle and weight loss with cancer risk and other chronic diseases are not fully understood, alterations in hormone signalling, especially insulin, sex steroids and adipokine pathways, may play an important role [10, 11].

Adipose tissue cells secrete several metabolically active peptides and proteins, amongst which are adiponectin and leptin. Adiponectin is an insulin-sensitizing, anti-angiogenic, anti-inflammatory hormone that plays a central role in energy homoeostasis, as well as lipid and glucose metabolism [12, 13]. The serum concentrations of adiponectin have been described as a positive predictor of all-cause cardiovascular mortality in patients with type 1 diabetes and could potentially be implemented as a biomarker for diabetes [14, 15]. Leptin is a key regulator of appetite, food intake and body weight. Leptin is also an important factor in energy homoeostasis, metabolism and adiposity [16, 17].

The effects of long-term (≥1 year) dietary weight loss in postmenopausal women by diet alone and exercise or diet and exercise combined on adiponectin and leptin levels have yet to be clearly established. The serum leptin concentration increases with weight gain and decreases with short-term diet-induced weight loss [18, 19]. The adiponectin concentration is inversely correlated with adiposity, although the impact of weight loss is unclear [20, 21].

The role of exercise on adipokines is also unclear. In a randomized controlled trial involving 320 postmenopausal women, a year-long aerobic-exercise intervention (225 min per week) was associated with increased adiponectin and decreased leptin concentrations [22]; however, another study showed that the effect of a hypocaloric diet and/or aerobic exercise in 40 overweight and obese postmenopausal women did not result in a change in adiponectin concentrations with a moderate weight-reduction programme or exercise regimen [9].

The present randomized controlled trial investigated the independent and combined effects of a reduced calorie diet and a moderate-to-vigorous aerobic-exercise intervention on circulating adiponectin and leptin concentrations in postmenopausal women. Our hypothesis was that a reduced calorie, weight-loss diet with moderate-to-vigorous aerobic exercise would produce a greater increase in the adiponectin level and a greater reduction in the leptin concentration than either intervention alone and compared with no lifestyle change (control). We hypothesized that the intervention effects on adiponectin and leptin levels would be stronger in women who were obese at baseline with a greater degree of weight or body fat loss during the intervention.

Subjects and methods

Study design and participants

The Nutrition and Exercise for Women (NEW) study was a 12-month randomized controlled trial, conducted between 2005 and 2009, which tested reduced calorie dietary weight loss and exercise effects on biomarkers of postmenopausal breast cancer risk and other end-points [23, 24]. Participants were recruited from the greater Seattle area (WA, USA) through mass mailings, media placements and community outreach.

Inclusion criteria were as follows: 50–75 years of age; BMI ≥25.0 kg m−2 (if Asian-American, ≥23.0 kg m−2); <100 min of moderate activity per week; postmenopausal; no menopausal hormone therapy for the past 3 months; no history of breast cancer, heart disease, diabetes mellitus or other serious medical conditions; fasting glucose <7.0 mmol L−1 and not taking diabetic medications; nonsmoking; alcohol intake of <2 drinks per day; ability to attend diet or exercise sessions at the intervention site; and a normal exercise tolerance test.

The trial design and recruitment (Fig. 1) have been previously reported [23]. Briefly, a total of 439 eligible women were stratified according to BMI (<30 kg m−2 or ≥30 kg m−2) and race per ethnicity (non-Hispanic White, Black, other), then randomized into one of four groups: (i) reduced calorie, weight-loss diet (diet; N = 118), (ii) moderate-to-vigorous intensity aerobic exercise (exercise; N = 117), (iii) combined reduced calorie, weight-loss diet and moderate-to-vigorous intensity aerobic exercise (diet + exercise; N = 117), and (iv) no diet or exercise change (control; N = 87).

Figure 1.

CONSORT diagram of the Nutrition and Exercise for Women (NEW) trial.

We used permuted-block randomization (ratio 0.75 : 1 : 1 : 1) to assign a proportionally smaller number of women to the control group. The Fred Hutchinson Cancer Research Center (FHCRC) Institutional Review Board in Seattle approved the study, and all participants signed informed consent.

Interventions and control group

The reduced calorie, weight-loss intervention was a modification of the Diabetes Prevention Program [25] and the Look AHEAD [26] lifestyle interventions and has been previously described [23]. The diet had a total energy intake goal of 1200–2000 kcal per day based on weight and <30% daily energy intake from fat. The weight-loss goal was 10% by 6 months, with maintenance thereafter. Dietitians with training in behaviour modification conducted sessions. Participants met individually with a study dietitian at least twice and attended weekly dietitian-led group meetings (5–10 women) in months 1–6. In months 7–12, participants attended monthly group meetings, in addition to phone or email contact with the dietitians. The diet + exercise group diet sessions were held separately from those of the diet group and participants were requested not to discuss diet during facility drop-in exercise training sessions.

The exercise intervention goal was 45 min of moderate-to-vigorous intensity exercise 5 days per week for 12 months. Participants attended three supervised sessions per week at the facility and exercised 2 days per week at home. The participants gradually increased exercise training to 70–85% of the observed maximal heart rate by the 7th week and maintained that intensity of exercise for the duration of the study.

Women randomized to the control group were asked not to change their diet or exercise habits. After 12 months, women were offered four weight-loss classes and 8 weeks of facility exercise training.

Measures

The primary outcomes for the present analyses were serum concentrations of adiponectin and leptin. We collected 12-hour fasting serum at baseline and 12 months. The samples were processed within 1 hour and stored at −70 °C. Laboratory assays were performed at the Northwest Lipid Research Laboratory at the University of Washington on 438 baseline samples (one blood sample was missing) and 399 (see below) 12-month follow-up samples.

Adiponectin was measured by a commercially available radioimmunoassay (Millipore Inc., Billerica, MA, USA) using 125I-labelled murine adiponectin as a control and a multispecies anti-adiponectin antibody. The assay has a sensitivity of 1 ng mL−1. Serum analyses were performed in duplicate and the results reported as the mean of the two values. The intra- and interassay coefficients of variation (CV) were 8.4% and 9.8%, respectively.

Leptin assays were performed with a commercially available radioimmunoassay (Millipore Inc.) using125I-labelled human leptin and a human leptin antiserum to determine the level of leptin by the double antibody/PEG technique. The sensitivity of this assay is 0.5 ng mL−1. Each sample was analysed in duplicate. The intra- and interassay CV were 9.1% and 14.3%, respectively.

We assessed demographics, medication use, lifestyle behaviours and anthropometrics at baseline and 12 months. Height and weight were measured with a stadiometer and standard scale, respectively, and BMI was calculated as kg m−2. Body fat was measured using a DXA whole-body scanner (GE Lunar, Madison, WI, USA).

Statistical analysis

Descriptive data are presented as the mean ± SD. Blood measures were log-transformed to compensate for the skewed distribution in the original measures. Accordingly, geometric means with 95% confidence intervals (CI) are presented. We examined the intervention effects on an intent-to-treat basis. The mean 12-month changes in the diet, exercise and diet + exercise groups were compared with controls and with the other intervention groups using the generalized estimating equations (GEE) approach to random-effects regression; this accounted for the correlation within each individual over time [23]. We used the Bonferroni correction (two-sided alpha = 0.05/6; critical P-value of P < 0.0083) to adjust for multiple comparisons.

We assessed changes in adipokines for each intervention group using GEE models by preplanned subgroups: (i) BMI (obese versus nonobese: <30 and ≥ 30 kg m−2) [27], (ii) weight loss (using cut points previously shown to reduce metabolic risk factors for diseases, such as diabetes: <5%, 5–10%, and ≥10%) [28], and (iii) body fat loss (tertiles, ≤2.58%, 2.58–6.35%, and >6.35%). For the subgroup analyses, P < 0.05 was considered statistically significant. We used sas software (version 9.1; SAS Institute, Cary, NC, USA) to perform all statistical analyses.

Results

The mean age of the participants was 58.0 ± 5.0 years. The majority of participants were non-Hispanic whites and highly educated (60% had a college degree or higher; Table 1). Three hundred ninety-nine (91%) of the randomized participants returned for a 12-month blood draw; there were no significant differences across the intervention groups (Fig. 1).

Table 1. Baseline characteristics of study participants (N = 438)
 Control (= 87)Diet (= 118)Diet + Exercise (= 116)Exercise (= 117)
Age, mean (SD)57.4 (4.4)58.1 (5.9)58.0 (4.4)58.1 (5.0)
Ethnicity No. (%)
Non-Hispanic white74 (85.1)101 (85.6)99 (85.3)98 (83.8)
African American6 (6.9)9 (7.6)15 (12.8)5 (4.3)
Asian/Pacific Islander2 (2.3)2 (1.7)2 (1.7)2 (1.7)
Hispanic/Latino3 (3.4)2 (1.7)2 (1.7)5 (4.3)
Other2 (2.3)4 (3.4)0 (0.0)5 (4.3)
Education
College degree, No. (%)59 (67.8)76 (64.4)81 (69.8)70 (59.8)
Marital status
Married or have partner, No. (%)59 (67.8)79 (67.0)69 (60.0)71 (60.7)
Smoking
Ever smoked, No. (%)32 (36.8)55 (46.6)47 (40.5)47 (40.2)
Anthropometrics
BMI, mean (SD), kg m−230.7 (3.9)31.0 (3.9)31.0 (4.3)30.7 (3.7)
Body fat, mean (SD), %47.8 (4.5)47.6 (4.4)47.9 (4.6)47.9 (4.1)
Body fat mass (SD), kg40.1 (8.5)39.8 (8.1)39.9 (8.2)39.4 (7.9)
Waist circumference, mean (SD), cm94.8 (10.2)94.6 (10.2)93.7 (9.9)95.1 (10.1)
Lifestyle behaviours
Aerobic fitness, mean (SD), mLkg−1min−123.1 (4.1)22.6 (3.8)23.6 (4.1)22.5 (4.1)
Physical activity, mean (SD), min per week23.8 (41.2)33.6 (45.5)32.4 (42.9)37.7 (43.7)
Total calorie intake, mean (SD), kcal per day1988 (669)1884 (661)1894 (639)1986 (589)
Per cent calorie intake from fat, mean (SD), %35.6 (6.9)33.1 (6.3)35.3 (7.3)33.6 (6.9)

Approximately, 42% (N = 49) of women assigned to the diet group and 60% (N = 69) of those in the diet + exercise group achieved the goal of a 10% reduction in weight over the 12-month period. The diet+exercise group completed 80% of the exercise goal (225 min per week), and the exercise group completed 85% of the exercise goal (data not shown).

At the end of the 12-month intervention period, there were improvements in all measures of adiposity in the three intervention groups compared with the control group. The most marked changes were in the diet + exercise group [23]; specifically, weight decreased by 10.8% (P < 0.01), whereas the reductions in weight in the diet and exercise groups were 8.5% (P < 0.01) and 2.4% (P = 0.08), respectively. Compared with controls, the largest decrease in body fat was observed in the diet + exercise group (−5.9%), the least reduction in body fat occurred in the exercise group (−1.7%) and the diet group lost −4.3% body fat (all P < 0.021).

The diet + exercise (6.6%) and diet groups (9.5%) had statistically significant increases in the mean serum adiponectin concentrations compared with controls (P < 0.0001; Table 2). In contrast, the mean adiponectin concentrations decreased by 2.6% and 3.3%, respectively, amongst women in the exercise and control groups. In the diet + exercise group, there was a mean 40% decrease in the leptin concentration compared with the control group (P < 0.0001). Statistically significant reductions in leptin concentrations also occurred in the diet (27%, P < 0.0001) and exercise groups (13%, P = 0.005). The difference in leptin level change between the diet and diet + exercise intervention groups was statistically significant (P = 0.0003); this was not the case for adiponectin (P = 0.32).

Table 2. Effect of individual and combined dietary weight loss and exercise interventions on adiponectin and leptin (geometric mean)
 Baseline12-Months ∆ from Baseline to 12-Months P value
N Mean (95% CI) N Mean (95% CI)∆ (%)
  1. GEE model, Last observation carried forward for 12-month missing values overall = 40 (8 control, 13 diet, 11 exercise, 8 diet + exercise), P < 0.05/6 = 0.0083 considered statistically significant.

  2. Pvalues comparing changes in adiponectin and leptin (adjusting for baseline values) versus control.

Adiponectin (μg mL−1)
Diet + Exercise116 12.8 (11.7, 13.9)11613.6 (12.5, 14.8)+0.8 (6.6) 0.0001
Diet118 12.4 (11.3, 13.5)11813.5 (12.5, 14.6)+1.2 (9.5) <0.0001
Exercise117 12.5 (11.5, 13.5)11712.1 (11.1, 13.2)−0.3 (−2.6)0.82
Control 87 12.8 (11.7, 14.0) 8712.4 (11.3, 13.5)−0.4 (−3.3)Ref.
Leptin (ng mL−1)
Diet + Exercise11623.8 (22.0, 25.6)11614.2 (12.8, 15.8)−9.5 (−40.1) <0.0001
Diet11823.1 (21.4, 24.9)11816.8 (15.2, 18.6)−6.3 (−27.1) <0.0001
Exercise11723.5 (21.7, 25.4)11720.5 (18.8, 22.4)−3.0 (−12.7) 0.005
Control 8724.9 (22.9, 26.9) 8724.4 (22.2, 26.8)−0.5 (−1.9)Ref.

Table 3 shows analyses stratified by BMI categories (<30 kg m−2 and ≥30 kg m−2). Amongst women randomized to the diet intervention, the serum adiponectin concentration increased in both BMI categories. Although the increase was more pronounced in participants with a BMI <30 kg m−2 (14.1% for a BMI < 30 kg m−2 vs. 3.3% for a BMI ≥ 30 kg/m−2), the difference between the two groups was not statistically significant. The leptin concentration decreased in both BMI categories in the diet and diet + exercise groups and to a lesser extent in the exercise group.

Table 3. Effect of individual and combined dietary weight loss and exercise interventions on adiponectin and leptin, stratified by baseline BMI (geometric mean)
 ControlDiet
Baseline12-month% Baseline12-month% P *
N Mean(LL–UL) N Mean(LL–UL) N Mean(LL–UL) N Mean(LL–UL)
Adiponectin (μg mL−1)
BMI < 30 kg m−24114.1(12.5–15.9)3714.2(12.6–16.0)+0.1+0.9 5512.8(11.2–14.5)5314.6(13.0–16.4)+1.8+14.1 0.001
BMI ≥ 30 kg m−24611.7(10.4-13.2)4211.1(9.7–12.6)−0.7−5.6 6312.0(10.8–13.4)5212.4(11.1–13.9)+0.4+3.3 0.002
      ** = 0.59
Leptin (ng mL−1)
BMI < 30 kg m−24119.8(17.8–22.0)3719.4(16.8–22.4)−0.4−1.9 5518.1(16.5–20.0)5312.0(10.5–13.6)−6.2−34.1 <0.0001
BMI ≥ 30 kg m−24630.4(28.0–33.0)4230.4(27.1–34.0)−0.1−0.2 6328.5(26.2–30.9)5220.7(18.4–23.4)−7.7−27.1 <0.0001
      ** = 0.31
 ExerciseDiet + Exercise
Baseline12-month%P *Baseline12-month%P *
N Mean(LL–UL) N Mean(LL–UL) N Mean(LL–UL) N Mean(LL–UL)
  1. For these subgroup analyses, P < 0.05 was considered statistically significant. LL, lower limit of 95% confidence interval; UL, Upper limit of 95% confidence interval; P-inter, P-value for interaction.

  2. *Pvalue comparing change from baseline to 12-months follow-up in intervention group versus controls within BMI strata. **Pvalue comparing difference in change from baseline to 12 months in intervention group versus controls in high BMI strata versus low BMI strata.

Adiponectin (μg mL−1)
BMI < 30 kg m−25713.6(11.9–15.5)5313.3(11.6–15.3)−0.3−1.90.995614.8(13.4–16.3)5315.9(14.5-17.5)1.1+7.70.046
BMI ≥ 30 kg m−26011.5(10.4–12.7)5311.0(9.8–12.3)−0.5−4.50.856011.1(9.8–12.6)5512.3(10.8–13.9)1.1+10.2 <0.0001
   P-inter = 0.85  P ** = 0.68
Leptin (ng mL−1)
BMI < 30 kg m−25718.6(16.8–20.6)5316.2(14.3–18.3)−2.4−3.00.095618.8(17.0–20.7)5310.2(8.9–11.7)−8.6−45.6 <0.0001
BMI ≥ 30 kg m−26029.4(26.9–32.0)5325.6(23.0–28.5)−3.8−2.9 0.02 6029.6(27.2–32.2)5518.4(16.3–20.8)−11.2−37.8 <0.0001
   P-inter = 0.89  ** = 0.28

When stratified by per cent weight loss (Table 4), women in the diet + exercise and diet groups with the greatest weight loss (>10%) had statistically significant increases in adiponectin (+11.7%; P = 0.047 and +18.5%; P = 0.02, respectively) compared with controls. The amount of weight loss from baseline was directly related to the increase in adiponectin (diet, P-trend < 0.002; diet + exercise, P-trend < 0.005) and the reduction in leptin across the weight-loss categories (diet, P-trend < 0.0001; diet + exercise, P-trend < 0.0001).

Table 4. Effect of individual and combined dietary weight loss and exercise interventions on adiponectin and leptin, stratified by percent weight loss (geometric mean)
 ControlDiet
Baseline12-month%Baseline12-month% P *
N Mean(LL–UL) N Mean(LL–UL) N Mean(LL–UL) N Mean(LL–UL)
Adiponectin (μg mL−1)
Weight loss < 5%6712.9(11.6–14.2)6612.2(11.1–13.5)−0.6−4.92810.8(9.4–12.5)2810.7(9.4–12.3)−0.1−0.80.18
Weight loss 5–10%913.7(10.3–18.2)915.3(11.4–20.5)+1.6+11.32713.4(11.3–15.9)2714.9(12.8–17.4)+1.5+11.00.96
Weight loss ≥10%410.7(7.6–15.2)410.3(7.8–13.6)−0.4−3.74912.1(10.5–13.9)4914.3(12.7–16.1)+2.2+18.5 0.02
      

P***=0.62; P**** = 0.02

P -trend = 0.0002

Leptin (ng mL−1)
Weight loss < 5%6725.0(22.8–27.4)6625.8(23.2–28.7)+0.8+3.32824.0(21.2–27.1)2824.1(21.5–27.1)+0.1+0.60.64
Weight loss 5–10%922.8(16.9–30.8)919.5(15.1–25.3)−3.3−14.62722.6(19.9–25.7)2717.1(14.7–19.9)−5.5−24.30.12
Weight loss ≥ 10%432.3(25.6–40.7)419.6(8.7–44.5)−12.7−39.34921.9(19.2–24.9)4911.9(10.2–13.7)−10.0−45.80.67
      

P*** = 0.29;P****=0.60

P -trend ≤0.0001

 ExerciseDiet + Exercise
Baseline12-month%P *Baseline12-month%P *
N Mean(LL–UL) N Mean(LL–UL) N Mean(LL–UL) N Mean(LL–UL)
  1. For these subgroup analyses, P < 0.05 was considered statistically significant.LL, lower limit of 95% confidence interval; UL, Upper limit of 95% confidence interval; P-trend, trend test of baseline to 12 month difference across subgroups.

  2. *Pvalue comparing change from baseline to 12-months follow-up in intervention group versus controls within weight-loss strata. ***Pvalue comparing difference in change from baseline to 12 months in intervention group versus controls in middle weight-loss strata versus low weight-loss strata. ****Pvalue comparing difference in change from baseline to 12 months in intervention group versus controls in high weight-loss strata versus low weight-loss strata.

Adiponectin (μg mL−1)
Weight loss < 5%7512.7(11.4–14.2)7512.2(10.8–13.7)−0.6−4.50.801811.9(10.3–13.8)1811.5(9.9–13.5)−0.4−3.40.59
Weight loss 5–10%2611.3(9.6-13.3)2611.6(9.8–13.7)+0.2+2.20.142111.0(8.7–13.9)2111.2(8.9–14.1)+0.2+2.10.12
Weight loss ≥ 10%413.4(12.0–15.1)412.8(10.2–15.9)−0.7−4.90.946914.0(12.7–15.5)6915.6(14.3–17.1)+1.6+11.70.07
   P*** = 0.10; P**** = 0.98  

P*** = 0.07;P**** = 0.047;

P -trend = 0.0005

Leptin (ng mL−1)
Weight loss < 5%7523.1(21.0–25.5)7521.9(19.8–24.2)−1.3−5.40.0471827.4(23.0–32.5)1828.0(23.9–32.8)+0.6+2.20.93
Weight loss 5–10%2624.6(20.6–29.5)2617.5(14.4–21.3)−7.10.0452121.4(17.9–25.7)2115.0(11.9–19.0)−6.4−29.9 0.01
Weight loss ≥ 10%421.7(12.4–37.9)411.0(6.3–19.1)−10.70.566923.9(21.6–26.4)6911.2(10.0–12.4)−12.7−53.20.33
   P*** = 0.43; P**** = 0.63  

P*** = 0.03; P**** = 0.26;

P -trend ≤0.0001

Stratification by body fat loss revealed similar patterns (Fig. 2 and Table S1) of statistically significant changes in adiponectin and leptin concentrations amongst participants in the intervention groups compared with controls. Because four levels of fat loss were investigated, and the control group showed few body fat changes, the comparisons were made in relation to the entire control group. In all three intervention groups, women with higher per cent body fat changes experienced greater mean changes in adiponectin and leptin concentrations. Trends towards increasing adiponectin levels amongst participants losing more fat were highly significant for participants in the diet group (P-trend < 0.0001). Reductions from 47.1% (exercise) to 55.3% (diet + exercise; both P = 0.0001) were observed amongst participants who lost >6.35% body fat compared with controls.

Figure 2.

Effect of individual and combined dietary weight loss and exercise interventions on adiponectin and leptin, stratified by change in % body fat (geometric mean).

An evaluation of the effects of the interventions on adipokines stratified by BMI is presented in Table S2.

Discussion

This randomized controlled trial amongst healthy overweight and obese postmenopausal women compared the individual and combined effects of a 12-month reduced calorie dietary weight-loss intervention, with or without exercise, on serum adiponectin and leptin concentrations. We observed substantial reductions in leptin concentrations in the three intervention groups, with the largest reduction (40%) amongst women in the diet + exercise group. Adiponectin concentrations increased amongst women in the diet + exercise and diet groups, but not amongst women in the exercise group. Regardless of the intervention, weight loss had a dose-dependent effect on leptin and adiponectin levels. The greatest increase in adiponectin levels and decrease in leptin levels occurred in participants in the diet and diet + exercise groups, who lost ≥10% of baseline weight or > 6.35% of body fat. The largest increase in adiponectin level approached 20% (diet group) and the largest decrease in leptin level was > 55% (diet + exercise group).

To date, randomized trials have tested the effect of weight loss on adiponectin and leptin levels, using hypocaloric diets and diets that differ in macronutrient composition [29, 30]. Summer et al. [29] reported that the adiponectin level increased with a low-carbohydrate, but not a low-fat, diet, with no correlation between weight loss and increase in adiponectin concentrations. In a 6-month weight-loss intervention study with dietary weight loss and resistance training in 46 healthy postmenopausal women, Drapeau et al. [31, 32] reported that a 7.7% weight loss in the caloric restriction group led to a 4.9% increase in adiponectin, whilst an 8.2% weight loss in the combined group led to a 12.8% rise in the serum adiponectin level.

The literature varies on the effect of exercise alone on adiponectin and leptin levels. Several randomized controlled trials have shown no relationship between exercise and changes in adiponectin concentrations, although there is some evidence that moderate- or high-intensity resistance training produces body composition changes that reduce circulating adiponectin concentrations [33]. Friedenreich et al. [22] reported that a year-long aerobic-exercise intervention (225 min per week) led to an 18.9% reduction in leptin levels in 320 previously inactive postmenopausal women, but no differences in adiponectin concentrations existed between exercisers and controls. In the current study, exercise alone had no effect on the adiponectin level and there was a modest effect (~13% reduction) on the leptin level. In an identical exercise intervention to that of the current study, Frank et al. [34] reported that 12 months of moderate intensity aerobic activity, 5 days per week for 45 min, resulted in a 7% reduction in leptin concentrations in postmenopausal women.

The results from our study are similar to the results reported by Christiansen et al. [30], in which 79 obese men and women were randomized to a hypocaloric diet with or without aerobic exercise training using a 3-group trial (diet, diet + exercise, and exercise). In that study, the diet and diet + exercise groups lost a mean of 12.3 kg and the exercise group lost a mean of 3.5 kg over 12 weeks. There was a statistically significant increase in serum adiponectin concentrations in the diet and diet + exercise interventions, but not in the exercise group [30]. In addition, the Diabetes Prevention Programme (DPP), which included a lifestyle group that prescribed a reduced calorie weight-loss programme and increased physical activity, showed an increase in adiponectin after 1 year in the lifestyle (diet + exercise) group [35]. Our diet intervention was based on the DPP lifestyle intervention, and had outcomes consistent with the DPP findings. Overall, our data and the data from other studies suggest that weight loss and/or reduction in body fat might be triggers for increases in adiponectin concentrations.

Our large sample size enabled us to investigate intervention effects stratified by several measures of body composition. We hypothesized that we would observe greater effects on biomarkers in obese women and women with greater abdominal adiposity. For adiponectin, women with a BMI ≥ 30 kg m−2 benefited most from a diet or diet + exercise intervention with respect to adipokine concentrations. We hypothesized that changes in weight and body fat would substantially modify the effects on adipokine concentrations. Changes in body composition influenced leptin independent of the intervention group. The results stratified by the proportion of weight and body fat loss showed strong trends from lesser to greater weight loss or body fat loss in the diet and diet + exercise intervention groups. These results are consistent with our hypothesis that greater reductions in weight, and specifically body fat, are needed to affect adiponectin and leptin concentrations.

Our study has broad implications with respect to the risk of cancer and other metabolic diseases. The most recent report from the World Cancer Research Fund stated that the link between obesity and cancer is convincing for cancers of the breast in postmenopausal women, colorectum, endometrium, kidney, pancreas and oesophagus [5]. Maintaining a healthy body weight, increasing physical activity and limiting energy-dense foods and drinks are some of the recommendations to reduce the risk of cancer [5]. Adipose tissue produces adiponectin and leptin, which are thought to mediate the effects of obesity on the risk of cancer, partially by influencing inflammatory and immune responses [36].

A number of recent prospective studies have reported that adiponectin and leptin may play a role in carcinogenesis as follows: the Nurses' Health Study reported an inverse association between adiponectin concentrations and postmenopausal breast cancer risk [37]; a nested case–control study involving 18 225 men showed that low serum adiponectin concentrations are associated with a higher risk of colorectal cancer [38]; and data from a cohort in Norway showed a nearly threefold increased risk of colon cancer amongst people with high leptin concentrations independent of BMI [39]. A systematic review of adiponectin and cancer concluded that measurement of adiponectin might serve as a useful means for predicting risk of obesity-related cancers [40]. Additionally, there is a defined relationship between insulin sensitivity, fasting insulinaemia and plasma adiponectin [41, 42]. A meta-analysis of 13 prospective studies showed an inverse link between the incidence of type 2 diabetes and plasma adiponectin levels [43]. The risk of developing type 2 diabetes decreases with increasing adiponectin concentrations, potentially due to suppression of hepatic gluconeogenesis, stimulation of fatty acid oxidation and glucose uptake in skeletal muscle [43-45]. Furthermore, studies have shown that serum adiponectin may be used as a biomarker predicting all-cause and cardiovascular mortality in patients with type 2 diabetes [14]. To gain more detailed knowledge, additional research is needed in establishing standardized routine adiponectin measurement methods and expertise in interpretation of the results [15]. Associations between obesity and leptin with cardiovascular, endocrine and inflammatory processes have been described. Leptin is responsible for energy regulation and satiety, thus leptin is strongly correlated with body fat [16]. A reduction in elevated leptin concentrations in the circulation can lead to an improvement in blood lipid levels, blood pressure and insulin sensitivity [46, 47].

This current study is the first to examine the individual and combined effects of a 12-month dietary and aerobic-exercise intervention on circulating adiponectin and leptin concentrations in a large sample of overweight and obese postmenopausal women. The strengths of the study include the large study size, with sufficient power to assess long-term changes in adiponectin and leptin concentrations, a randomized controlled design with three intervention groups, high retention (91%) and high adherence to the intervention programmes. Another key strength of the study was that this approach can be instituted clinically or by well-motivated individuals and groups; the group-based modification of the DPP intervention can be easily adopted and the exercise component, which consisted primarily of brisk walking, is similarly applicable [25].

We recognize that there are some limitations to the current study. The generalizability of these results may be limited by the relatively homogeneous sample of postmenopausal women. The study population was primarily non-Hispanic white women, and the effects of weight loss or exercise on adipokines in women from other races or ethnic groups cannot be inferred without additional data. Furthermore, only one dietary weight-loss programme and one exercise programme were tested; therefore, we are unable to speculate on the effects of other weight-loss methods or other types or intensities of exercise programmes on adiponectin and leptin levels. Our results and various available diet and exercise intervention programmes need to be studied, additionally in different ethnicities, for potential implementation into clinical practice.

In conclusion, dietary weight-loss intervention, with or without exercise, increased the serum adiponectin concentration in overweight and obese postmenopausal women, whereas exercise alone showed limited beneficial effects. Substantial weight loss or per cent body fat loss due to dietary weight-loss interventions resulted in an increase in the adiponectin concentration and reduction in leptin of >50%. The results from our study can provide insight into the biologic mechanisms that mediate the effects of weight loss and exercise on cancer risk, as well as providing further data on the public health impact of relatively simple lifestyle interventions on diet and weight.

Financial disclosures

None.

Conflict of interest statement

None for Clare Abbenhardt, Catherine M. Alfano, Mark H. Wener4, Kristin L. Campbell, Catherine Duggan, Karen E. Foster-Schubert, Angela Kong, Adetunji T Toriola, John D. Potter, Caitlin Mason, Liren Xiao, George L. Blackburn, Carolyn Bain, Cornelia M. Ulrich Anne McTiernan: Consultant for Metagenics (minor); Merck stock (minor).

Acknowledgements

We would like to thank the study staff and the participants for their dedication to the study. This work was supported by the National Cancer Institute at the National Institutes of Health (grants U54-CA116847, R01 CA102504; 5KL2RR025015 to K.F.S; R25 CA94880 and R25 CA057699 to A.K.); and the Canadian Institutes of Health Research (Fellowships to K.L.C & C.M). None of the funding agencies were involved in the trial design or conduct. During the trial, Dr. Alfano was a faculty member at Ohio State University and relocated to the NCI following completion of her efforts on the NEW trial.

Ancillary