European Journal of Applied Physiology

, Volume 109, Issue 5, pp 823–828

Plasma adiponectin concentration is associated with the average accelerometer daily steps counts in healthy elderly females

Authors

    • Institute of Sport Pedagogy and Coaching Sciences, Centre of Behavioural and Health SciencesUniversity of Tartu
  • Tatjana Kums
    • Institute of Exercise Physiology and Physiotherapy, Centre of Behavioural and Health SciencesUniversity of Tartu
  • Toivo Jürimäe
    • Institute of Sport Pedagogy and Coaching Sciences, Centre of Behavioural and Health SciencesUniversity of Tartu
Original Article

DOI: 10.1007/s00421-010-1423-9

Cite this article as:
Jürimäe, J., Kums, T. & Jürimäe, T. Eur J Appl Physiol (2010) 109: 823. doi:10.1007/s00421-010-1423-9

Abstract

This study is aimed to evaluate whether circulating adiponectin concentration is associated with physical activity (PA) level in healthy older females. To date, daily PA in older adults (≥65 years) has primarily relied on self-report. This study used accelerometry, which objectively measured minute-by-minute movement to assess PA volume and intensity performed by elderly females. In addition, body composition, leptin and insulin resistance values were measured to assess the influence of these parameters on the possible relationship between adiponectin and PA levels in this specific age group of older women. On 49 women (mean age: 73.6 ± 4.2 years), adiponectin, leptin, insulin resistance, body composition and 7-day PA parameters were measured. Average daily accelerometer step counts and time spent in different PA levels were obtained from 7-day PA measurement. Average daily accelerometer step-count was 7,722 ± 3,069 steps day−1 and the recommended 150 min weekly of at least moderate/vigorous PA in bouts of at least 10 min was achieved by 71.4% (35/49) of the participants. Correlation analysis showed that plasma adiponectin concentration (16.0 ± 6.1 μg ml−1) was related (P < 0.001) to steps per day (r = 0.438) and leptin (r = −0.443) values. Multivariate regression analysis further revealed that only steps per day and leptin were independent predictors of circulating adiponectin concentration in healthy elderly females. In conclusion, these data support the hypothesis that being physically active is associated with better adiponectin concentration and a reduced risk of having metabolic disease risk in the specific group of healthy elderly females.

Keywords

AdiponectinPhysical activityElderly females

Introduction

Adiponectin is an adipocytokine that seems to be exclusively secreted by adipocytes and is the most abundant adipose tissue protein (Matsuzawa et al. 2004). Adiponectin correlates negatively with adiposity (Jürimäe and Jürimäe 2007; Jürimäe et al. 2009a, b). Lower circulating adiponectin concentrations relative to the normal controls have been observed in human subjects with obesity, type 2 diabetes mellitus or cardiovascular disease in several studies (Dekker et al. 2008; Jürimäe et al. 2009a; Matsuzawa et al. 2004). Different studies have also shown that adiponectin may improve the lipid profile (Cnop et al. 2003; Yamauchi et al. 2001). However, the entire spectrum of predictors of circulating adiponectin concentrations remains to be fully elucidated (Gavrila et al. 2003; Kizer et al. 2008). For example, age (Cnop et al. 2003; Jürimäe and Jürimäe 2007) and menopause transition (Isobe et al. 2005) may have an influence on circulating adiponectin concentration. It is well established that aging and also menopause transition have been characterized by changes in body composition, including a decrease in body fat free mass (FFM) and an increase in body fat mass (FM) (Jürimäe and Jürimäe 2007; Kanaley et al. 2001).

The age-related changes in body composition are caused by a reduction in specific anabolic hormones and also in physical activity (PA) level (Bertoli et al. 2006; Jürimäe et al. 2009c). A decreased total PA level has been associated with increased risk of having cardiovascular disease in older people (Aoyagi and Shephard 2009; Yu et al. 2009). While questionnaires mostly have been used to monitor PA in elderly, evidence of their precision and validity remains limited (Davis and Fox 2007; Harris et al. 2009). Most PA in older people is walking based and hence accelerometers should be used to measure PA in these subjects (Davis and Fox 2007; Harris et al. 2009). Accelerometers are sensitive to walking and objectively quantify PA as a continuous variable (Harris et al. 2009). Accelerometers have been validated in older people (Ekelund et al. 2002) and used successfully to describe PA levels in small groups of older people (Davis and Fox 2007; Ekelund et al.2002). To date, the first population-based sample of older people with objective PA level assessed by accelerometry showed that PA levels in these subjects are relatively low (Hagströmer et al. 2007; Harris et al. 2009). The Harris et al. (2009) study was also the first to examine the associations of objectively measured PA levels with physical health, disability, anthropometric measures and psychological factors.

To our knowledge, no investigations have been performed to examine the possible association of adiponectin with PA levels measured objectively using accelerometers in healthy older people. Accordingly, our main objectives were to assess customary PA levels in a specific sample of older women and to examine the association with circulating adiponectin concentration. In addition, body composition, leptin and insulin resistance values were measured to assess the influence of these parameters on the possible relationship between adiponectin and PA in this specific age group of older women. It was hypothesized that adiponectin is independently associated with total PA levels in healthy older women.

Methods

Participants

Forty-nine older women at least 65 years old participated in this study. The volunteers were recruited from Tartu, Estonia. They were taking part in the 60-min gymnastics lessons twice a week and had been doing so for at least the last 5 years. All participants signed an informed consent that was approved by the Medical Ethics Committee of the University of Tartu, Tartu, Estonia. Prior to study enrolment, volunteers completed medical and PA questionnaires. The exclusion criteria were: the presence of any disease causing significant impairment of the nutritional status, the presence of endocrine diseases, the occurrence of injury and treatment with special diets. All women were asked for two visits to complete the testing. On the first visit, women had anthropometric parameters measured and a venous blood sample was taken in the morning after a 10-h fast. The second measurement session consisted of body composition measurement by dual energy X-ray absorptiometry (DXA). Measurement sessions were separated by approximately 1 week depending on the participant’s schedule and DXA availability. In addition, PA was assessed by accelerometry during 1 week. Participants were asked to maintain their usual activites during the study period (Harris et al. 2009).

Anthropometric and body composition measurements

Height was measured using a Martin metal anthropometer to the nearest 0.1 cm with a standard technique. Body mass was measured with minimal clothing to the nearest 0.05 kg using a medical electronic scale (A&D Instruments, Oxfordshire, UK) and body mass index (BMI) was calculated as body mass (kg) divided by height (m2). Total body FM and FFM were measured by DXA using the DPX-IQ densitometer (Lunar Corporation, Madison, WI, USA) equipped with adult, proprietary software, version 3.6. Participants were scanned in light clothing while lying flat on their backs with arms at their sides. The standard participant positioning was used for total body measurements and analysed using the extended analysis option. The standard manufacturer’s skeletal landmarks were used to define trunk and leg fat. Body fat distribution was calculated as the ratio of trunk fat (kg) to leg fat (kg) (Jürimäe and Jürimäe 2006). Coefficients of variations (CVs) for measured FM and FFM parameters were <2%.

Physical activity assessment

Physical activity was measured by accelerometry using the Actigraph AM-7164 monitoring device (Actigraph, Ft. Walton Beach, FL, USA). The Actigraph was worn superior to the iliac crest in a custom pouch, secured to the participant’s belt by a fastening. Participants were asked to wear accelerometer for seven consecutive days during waking hours and remove it for sleep and bathing only. Activity was recorded using 5 s epochs. Participants also completed a log to record when the Actigraph was worn. Accelerometer traces were checked using the Actigraph-provided ActiLife Monitoring System, alongside activity logs. Participants with less than 5 days of data were excluded (Davis and Fox 2007; Harris et al. 2009). In addition, recorded data were viewed for signs of malfunction (i.e. unusually low or high counts and the continuous data with the same values). Data were assessed for outliers, identified cases were checked and files reviewed, and excluded if data were outside the range or pattern expected of normal PA (Davis and Fox 2007).

Each minute epoch was then assigned an activity level based on the number of counts per minute: sedentary PA (<260 counts), light PA (260–1,951 counts) or moderate/vigorous PA (≥1,952 counts) (Bassett et al. 2000; Janney et al. 2008). The cut-off point of 1,952 counts per min for moderate/vigorous PA has been used to define the boundary between light and moderate activity (Freedson et al. 1998). Average accelerometer daily step count (steps per day) was chosen as the main objective PA outcome as walking is the predominant PA in this age group and step counts give readily understandable effect estimates (Harris et al. 2009). Average daily activity counts (total activity) were also analysed (Harris et al.2009).

Blood sampling and analysis

A 10 ml blood sample was obtained from the antecubital vein with the participants in the upright position in the morning (0700–0800 hours) after an overnight fast. Plasma was separated and frozen at −20°C for later analysis. Total adiponectin concentration was determined in duplicate using a commercially available RIA kit (cat. No. HADP-61 HK; Linco Research, USA). The intra- and inter-assay CV values were <7%. Leptin concentration was also assessed in duplicate by RIA (Mediagnost GmbH, Reutlingen, Germany). This assay has the intra- and inter-assay CV values of less than 5%. The concentration of insulin was determined in duplicate on an Immulite 2000 (DPC, Los Angeles, CA, USA). The intra- and inter-assay CVs were 4.5 and 12.2%, respectively, at an insulin concentration of 6.6 μIU ml−1. Glucose concentration was measured using the hexokinase/glucose-6-phosphate dehydrogenase method with a commercial kit (Boehringer, Mannheim, Germany). The insulin resistance index from fasting plasma insulin and plasma glucose levels was estimated using the homeostasis model assessment (HOMA) = fasting plasma insulin (μIU ml−1) × fasting plasma glucose (mmol l−1)/22.5 (Matthews et al. 1985). The greater the HOMA value, the greater was the level of insulin resistance.

Statistical analysis

Statistical analyses were performed with SPSS for Windows (Chicago, IL, USA) and the means (±SD) were determined. Spearman correlation analysis was performed to assess bivariate relationships of plasma adiponectin concentration with other measured variables. Bivariate and multivariate regression analyses were also performed to evaluate potential associations of adiponectin with PA, body composition and metabolic variables. A P value of less than 0.001 represented statistical significance after adjusting for multiple analysis.

Results

The mean (±SD), minimum and maximum values of measured characteristics for study population are presented in Table 1. The recommended 150 min weekly of at least moderate/vigorous PA in bouts of at least 10 min was achieved by 71.4% (35/49) of the participants. However, all participants achieved some moderate/vigorous PA during a week. Correlation analysis revealed that plasma adiponectin concentration was related (P < 0.001) to steps per day (r = 0.438) and leptin (r = −0.443) values. All other correlations between adiponectin and measured parameters were not significant (r < 0.349, P > 0.001).
Table 1

Mean (±SD) subject characteristics of study population (n = 49)

Variable

Mean ± SD

Range

Age (years)

73.6 ± 4.2

67–81

Height (cm)

158.5 ± 5.1

148.3–171.8

Body mass (kg)

66.3 ± 9.9

46.3–88.1

BMI (kg m−2)

26.4 ± 3.8

18.9–29.9

FM (%)

35.2 ± 8.3

14.1–48.6

FM (kg)

23.5 ± 7.9

6.8–42.1

FFM (kg)

41.5 ± 2.8

33.7–47.5

Trunk fat (kg)

11.1 ± 3.9

2.7–18.5

Trunk fat:leg fat ratio

1.71 ± 0.37

1.07–2.38

Adiponectin (μg ml−1)

16.0 ± 6.1

7.5–29.8

Leptin (ng ml−1)

13.9 ± 6.8

5.2–27.1

Insulin (μIU ml−1)

6.8 ± 3.7

2.0–16.7

Glucose (mmol l−1)

5.1 ± 0.4

4.5–5.9

HOMA

1.73 ± 1.41

0.41–9.12

Moderate/vigorous PA (min day−1)

34.7 ± 23.1

3.8–83.0

Light PA (min day−1)

197.3 ± 65.2

86.0–327.2

Sedentary PA (min day−1)

547.4 ± 93.9

383.5–695.3

Total activity (counts day−1)

537,120 ± 220,320

187,200–1,339,200

Steps per day

7,722 ± 3,069

2,550–13,519

A strong association of plasma adiponectin with age was found (β = 0.931, P < 0.0001), which remained significant (P < 0.001) after controlling for BMI, FM, FFM, trunk fat, trunk fat:leg fat ratio, leptin, insulin or HOMA values. In addition, a strong association between plasma adiponectin concentration and parameters of overall obesity (BMI, FM), central obesity (trunk fat, trunk fat:leg fat ratio), PA (sedentary PA, light PA, moderate/vigorous PA, total activity, steps per day), leptin and insulin resistance (insulin, HOMA) was found (Table 2). The association of adiponectin with steps per day and leptin values remained significant after accounting for age separately, and for age and FM, age and trunk fat or age and HOMA together. In addition, negative associations between adiponectin with overall (BMI, FM) and central (trunk fat, trunk fat:leg fat ratio) obesity remained significant after controlling for age and HOMA.
Table 2

Bivariate and multivariate regression analyses of body composition, physical activity and metabolic factors as predictors of plasma adiponectin levels (n = 49)

Variable

β1*

β2

β3

β4

β5

BMI (kg m−2)

−0.914

−0.269

−0.237

−0.286

−0.535*

FM (kg)

−0.850

−0.299

−0.343

−0.365*

FFM (kg)

0.928

0.017

0.346

0.303

0.128

Trunk fat (kg)

−0.844

−0.295

−0.104

−0.407*

Trunk fat:leg fat ratio

−0.895

−0.318

−0.229

−0.574*

Moderate/vigorous PA (min day−1)

0.825

0.175

0.129

0.112

0.220

Light PA (min day−1)

0.913

0.305

0.257

0.232

0.319

Sedentary PA (min day−1)

0.896

0.462*

0.348

0.311

0.406*

Total activity (min day−1)

0.908

0.333

0.278

0.261

0.351*

Steps per day

0.923

0.434*

0.400*

0.397*

0.458*

Leptin (ng ml−1)

−0.767

−0.390*

−0.355*

−0.351*

0.371*

Insulin (μIU ml−1)

−0.801

−0.087

−0.057

−0.052

HOMA

−0.769

−0.014

−0.155

−0.168

β1 bivariate standardized linear regression coefficient, β2 multivariate standardized linear regression coefficient adjusted for age, β3 multivariate standardized linear regression coefficient adjusted for age and FM, β4 multivariate standardized linear regression coefficient adjusted for age and trunk fat, β5 multivariate standardized linear regression coefficient adjusted for age and HOMA

* Statistically significant, P < 0.001

Discussion

The present study investigated whether plasma adiponectin concentration is associated with different PA, body composition and metabolic variables in healthy older women. The results of our investigation indicated that plasma adiponectin levels were independently related to the average accelerometer daily step count (steps per day) and leptin values in a specific group of healthy older women. To our knowledge, this may be the first study reporting the independent relationship of circulating adiponectin concentration with the amount of daily walked steps in healthy older women.

The average accelerometer daily step counts in the older women of the present study (7,722 ± 3,069 steps day−1) are in accordance with a review suggesting 6,000–8,500 steps day−1 for healthy older adults (Tudor-Locke and Myers 2001) and higher compared to the recent population-based sample of older adults (Harris et al. 2009). However, the recommended 150 min weekly of at least moderate/vigorous PA level in bouts of at least 10 min was achieved by 71.4% (35/49) of the participants. In comparison, Davis and Fox (2007) and Harris et al. (2009) reported only 1.8% (3/163) and 2.5% (6/238) of participants achieving recommended PA level, respectively. The differences in PA levels between current study and other studies (Davis and Fox 2007; Harris et al. 2009) could be explained by the participation in organised gymnastics lessons by our seniors. Other studies have also highlighted the importance of organised activity programmes to increase PA level and health in older people (Bravata et al.2007; Talbot et al. 2003).

The increased PA level as described by the average accelerometer daily step counts was independently associated with circulating adiponectin levels in our specific group of healthy older women (see Table 2). Adiponectin concentrations (16.0 ± 6.1 μg ml−1) in our healthy seniors were similar with adiponectin levels obtained in older women without cardiovascular disease (Dekker et al. 2008; Kizer et al. 2008) and type 2 diabetes (Snijder et al. 2006) risk factors. Dekker et al. (2008) argued that older women with higher adiponectin levels also have higher participation level in various sport activities. In addition, significant associations between adiponectin concentrations and cardiovascular fitness parameters have been observed in middle-aged and older women (Jürimäe and Jürimäe 2007), while Hulver et al. (2002) suggested that high adiponectin level serves as a protective mechanism against the development of cardiovascular diseases. In contrast, older women with unfavourable measures of body composition and different disease risk factors present lower adiponectin concentrations (Dekker et al. 2008; Kizer et al. 2008; Snijder et al. 2006). Taken together, the results of present study suggest that higher PA level is important to prevent a decrease in circulating adiponectin concentration in older people.

Higher adiponectin concentration protects against metabolic diseases in older population (Dekker et al. 2008). Furthermore, Yamauchi and Kadowaki (2008) concluded that specifically high molecular weight (HMW) adiponectin plays a crucial role in obesity-linked insulin resistance and metabolic syndrome. It could be argued that the limitation of the present study is that total adiponectin and not different adiponectin oligomers were measured in our subjects. HMW adiponectin appears to be more active (Yamauchi and Kadowaki 2008) and therefore could better characterise metabolic diseases in older population. In contrast, Sattar et al. (2008) suggested that HMW adiponectin is not directly linked to metabolic diseases in older women. However, to our knowledge, no studies have yet investigated the effects of PA on adiponectin oligomers in older adults. Accordingly, further studies are needed to investigate the possible association between PA and adiponectin oligomers before any conclusion can be drawn.

Another finding of the present investigation was an independent association between plasma adiponectin and leptin levels in our specific group of elderly females (see Table 2). According to our results, leptin is a negative predictor of circulating adiponectin concentration independent of body composition and insulin resistance values in normal weight healthy older women. This finding is similar to other studies with women at different ages (Cnop et al. 2003; Jürimäe and Jürimäe 2007; Jürimäe et al. 2009a, b; Ryan et al. 2003) and in contrast to the study by Gavrila et al. (2003) in pre- and postmenopausal women with a wide range of obesity levels. Gavrila et al. (2003) suggested that adiponectin and leptin may represent two different and independent pathways that control insulin sensitivity. However, the different results of present study with a relatively homogeneous group of healthy elderly females and that of Gavrila et al. (2003) study could be explained by the differences in overall adiposity values and/or fat distribution characteristics. In accordance with this, our recent study with middle-aged premenopausal women (Jürimäe et al. 2009a) found an independent association of adiponectin with the central fat distribution index (i.e. trunk fat–leg fat ratio) in normal weight but not in overweight women. Adiponectin was associated with overall and central adiposity values also in healthy elderly women of the present investigation (see Table 2). Accordingly, the relationship between plasma adiponectin concentration and different adiposity values may depend on the degree of adiposity, which in turn may modulate the association between adiponectin and leptin. However, similar to the results of the present investigation, Huypens (2007) suggested that adiponectin production is controlled in part by the hypothalamic action of leptin. In addition, it has been proposed that adipocyte-generated endocrine signals, such as adiponectin and leptin, control systemic insulin sensitivity as a part of a broader control mechanism of energy balance (Huypens 2007). An association between adiponectin and insulin resistance measures was also found in our subjects (see Table 2).

In conclusion, these data support the hypothesis that being physically active is associated with higher adiponectin concentrations and a reduced risk of metabolic disease in older women. Results of the present investigation demonstrated an independent association between circulating adiponectin levels and the average accelerometer daily step counts in healthy older women.

Acknowledgments

This study was supported by Estonian Science Foundation Grant GKKSP 6638.

Copyright information

© Springer-Verlag 2010