Relation of dietary and other lifestyle traits to difference in serum adiponectin concentration of Japanese in Japan and Hawaii: the INTERLIPID Study

Abstract

Background:

In initial analyses, we found significant differences between serum adiponectin in Japanese living in Japan and Hawaii.

Objective:

We investigated whether differences in dietary and other lifestyle factors explain higher serum adiponectin concentrations in Japanese in Japan compared with Japanese emigrants living a Western lifestyle in Hawaii.

Design:

Serum adiponectin and nutrient intakes were examined by standardized methods in men and women aged 40–59 y from 2 population samples, one Japanese American in Hawaii (99 men, 104 women), the other Japanese in central Japan (124 men, 125 women). Multiple linear regression models were used to assess the role of dietary and other lifestyle traits in accounting for serum adiponectin difference between Hawaii and Japan.

Results:

Mean adiponectin was significantly higher in Japan than in Hawaii(10.5 ± 5.5 μg/mL compared with 6.7 ± 3.2 μg/mL in men, P = 0.002; 12.9 ± 5.9 μg/mL compared with 9.4 ± 4.2 in women, P < 0.0001). In men, the difference in body mass index (BMI; in kg/m²) in the 2 populations explained ≈90% of the difference in serum adiponectin; in women, only 29%. In multiple linear regression analyses in women further adjustment for physical activity and 4 nutrients (nonvegetable protein, n–3 polyunsaturated fatty acid, arachidonic acid, dietary cholesterol) produced a further reduction in the coefficient for the difference (total 56%); P value for the difference became 0.074.

Conclusions:

The significantly highermean serum adiponectin concentration in Japan than in Hawaii may be attributable largely to differences in BMI. Differences in nutrient intake in the 2 samples were associated with only modest relation to the adiponectin difference.

INTRODUCTION

Adiponectin is the most abundant adipose-specific protein. Its expression is reduced in obesity, insulin resistance, and type 2 diabetes; plasma concentrations are inversely related to body weight, especially visceral adiposity (1–3). Recent research indicates that adiponectin may have antiinflammatory, antiatherogenic, and antidiabetic properties (4, 5). Lower serum adiponectin concentrations are reported to be associated with coronary artery disease (6–8).

In the INTERLIPID Study, we hypothesized that dietary or lifestyle factors may affect serum adiponectin independent of body mass index (BMI; in kg/m²). INTERLIPID, an ancillary study of the International Study of Macro/micronutrients and Blood Pressure (INTERMAP), investigated risk factors for coronary heart disease in 4 Japanese population samples in Japan and a Japanese American population sample in Hawaii (9, 10). In INTERMAP, dietary surveys were conducted with a highly standardized protocol in 17 random population samples in 4 countries (Japan, People’s Republic of China, the United Kingdom, and the United States) (11, 12).

In our initial INTERLIPID data analyses, we found significant sex-specific differences in serum adiponectin concentrations of middle-aged persons from community-based Japanese samples living in Japan and Hawaii. We hypothesized that these observed differences are largely related to differences in dietary and non-dietary lifestyle factors between the 2 population samples; data from previous studies show that Japanese in Japan and Japanese Americans in Hawaii have significantly different dietary patterns as well as BMI (9–11).

SUBJECTS AND METHODS

Detailed methods of the INTERMAP Study were described previously (11, 12). They are summarized here. Two standardized blood pressure measurements were made on each of 4 different days; medical and lifestyle information, 4 in-depth multi-pass 24-h dietary recalls, and 2 timed 24-h urine collections were obtained for each participant. In addition, nonfasting blood was drawn from INTERLIPID participants (9, 10). We used data on analytes measured in those samples, as well as INTERMAP dietary and other data.

Participants

INTERLIPID participants aged 40–59 y were from 5 INTERMAP research centers: 4 in Japan and 1 in Hawaii (9,10). For the present study, serum adiponectin concentrations were measured in 2 samples, one from Japan and one from Hawaii. The 2 populations samples were 1) Japanese residents in Aito Town, arural town in Shiga prefecture, central Japan (129 men and 129 women) and 2) third- and fourth-generation offspring of Japanese emigrants living in Honolulu, HI (100 men and 106 women). Participants in Honolulu were asked about the ethnicity of their mother and father; those included in the study responded 100% Japanese to both. Aito Town was chosen because it was the only Japanese community sample; the other 3 samples were of factory workers. Although Aito Town is a rural community, only ≈15% of the residents are engaged in primary industries; 43% are involved in secondary industries; 42% are involved in tertiary industries. They were not significantly different from the other 3 samples in Japan in terms of BMI. Only small differences were observed in lifestyle and dietary habits among the 4 samples in Japan; differences of those variables in Japan and Hawaii were larger (11). Among participants in these 2 samples, 12 persons (9 Japanese, 3 Japanese American) were excluded because their serum C-reactive protein (CRP) concentrations were >10 mg/L, leaving 249 persons (124 men and 125 women) in Japan, and 203 (99 men and 104 women) in Hawaii.

Ethics committees of the Shiga University of Medical Science, the Pacific Health Research Institute, and Northwestern University approved the study protocol. Written informed consent was obtained from all participants.

Anthropometric and lifestyle assessment

Participants visited the research centers 4 times on 2 pairs of consecutive days 3 wk apart on average. Height and weight with light clothes were measured at each visit. With the use of a questionnaire, trained observers inquired about physical activity, smoking status, previous medical history of cerebrovascular or cardiovascular diseases or diabetes, use of medication, and so forth. Hypertension was defined as systolic bloodpressure ≥ 140 mm Hg, diastolic blood pressure≥90 mm Hg, or use of antihy-pertensive medication. Diabetes mellitus was defined as glycated hemoglobin ≥ 6.5% or use of antidiabetic medication.

Dietary assessment

Four in-depth multipass 24-h dietary recalls per participant were conducted by specially trained dietary interviewers during the 4 visits. Before the data collection, a supervising nutritionist in each country trained all interviewers and certified that they had the appropriate skills to conduct dietary interviews and to handle dietary data with computers. Standardized on-going quality control procedures were adopted to ensure the quality of dietary data throughout data collection (12). Up-dated Standard Tables for Food Composition in Japan, 4th edition, with matched fatty acid values and micronutrients, were used to calculate Japanese nutrient intakes. In the United States, dietary assessment was performed with the use of the Nutrition Data System, Nutrition Coordinating Center, University of Minnesota. The Nutrition Coordinating Center, in cooperation with country nutritionists, was responsible for updated, standardized country-specific databases on nutrient composition (83 nutrients) of all foods consumed by INTERMAP participants and for assuring quality and comparability of these nutrient data (13). Although dietary records of the second and fourth day included survey days at research centers, participants were asked not to change their dietary habits. In fact, nutrient intakes and 24-h urinary sodium excretions were not different between odd and even days. All participants for the present study attended all 4 study visits; their energy intakes from all 24-h dietary recalls were between 500 and 5000 kcal/d.

Biochemical measurements

For the INTERLIPID Study, nonfasting blood was drawn on the second day of the first 2-day visit pair. Serum and plasma were obtained by centrifugation within 30 min of blood drawing and immediately refrigerated. Within 24-h, all specimens were frozen and stored locally at –70 °C. Samples from the Hawaiian and Japanese centers were shipped to a central laboratory in Japan on dry ice. Individual samples from the 2 centers were randomly allocated for analysis to avoid systematic measurement bias. The central laboratory was standardized by the Lipid Standardization Program, Centers for Disease Control and Prevention, Atlanta, GA; it successfully met the criteria of precision and accuracy of control (14). The laboratory is currently a member of the Cholesterol Reference Method Laboratory Network (13). Serum concentrations of total cholesterol, HDL cholesterol, and LDL cholesterol were directly measured by enzymatic methods on an autoanalyzer (Hitachi 7107; Hitachi, Tokyo, Japan). We included standard sera of known values with each batch; no significant differences were observed in the standard serum concentrations among the batches. The analytic CVs were <3% for total cholesterol, LDL cholesterol, HDL cholesterol, and triacylglycerols. Although significant postprandial increases in blood triacylglycerol concentrations occurs, postprandial stability of total cholesterol, HDL cholesterol, and LDL cholesterol were shown (15, 16). To assess postprandial stability of adiponectin, Peake et al (17) studied postprandial concentrations of adiponectin in 24 normal persons and 20 first-degree relatives of patients with type 2 diabetes; adiponectin postprandial concentrations were not significantly different from fasting concentrations, and no breakdown products of adiponectin were detectable in post-prandial samples by Western blotting. Pischon et al (18) found that human adiponectin concentrations were stable in whole blood stored in EDTA or sodium-heparin evacuated tubes when placed on ice packs and stored in Styrofoam containers for ≤36 h. Furthermore, with control for changes in BMI, blood adiponectin concentrations of persons did not significantly change during a period of 1 y; ie, there was high-level reproducibility and intraindividual stability.

Serum adiponectin concentrations were measured by an enzyme-linked immunosorbent assay kit (Otsuka Pharmaceutical Co, Ltd, Tokyo, Japan) at the central laboratory (19). CRP was measured by immunoturbidimetric assay at the central laboratory.

Data analyses

For each person, means of the individual nutrients from the four 24-h dietary recalls were used in the analyses. Data are presented as the contribution to total energy intake [in percentage of kcal (%kcal)] from total carbohydrates, total fat, monounsaturated fatty acids, saturated fatty acids (SFAs), polyunsaturated fatty acids (PUFAs), long chain n–3 PUFAs, α-linolenic acid, n–6 PUFAs, dietary cholesterol (in mg/1000 kcal), total dietary fiber (in g/1000 kcal), and alcohol (%kcal). In INTERMAP, intakes of arachidonic acid (ARA, n20:4), eicosapentaenoic acid (EPA, n20:5), docosahexaenoic (DHA, n22:6), and docosapentaenoic acid (DPA, n22:5) were estimated. ARA intake was included in the analyses because of its reported adverse, including atherogenic, effects. Intakes of EPA, DHA, and DPA were included because some components of n–3 PUFAs were not from marine foods; also, there were striking differences between Japanese in Japan and Hawaii in fish intake. The intakes of EPA, DHA, and DPA of the participants were highly correlated (Pearson partial correlation coefficients between any 2 of the 3 fatty acids > 0.9). Keys dietary lipid score, predictive of serum total cholesterol, was calculated as 1.35 × (2 SFA – PUFA) + 1.5 × C^½, where SFA is the percentage of total kilocalories from SFAs; PUFA, percentage from polyunsaturated fatty acids; C, dietary cholesterol in mg/1000 kcal (20). BMI was calculated. Sample average number of cigarettes smoked per day was calculated including nonsmokers. Student’s t tests were used to compare means of Japanese and Japanese Americans; chi-square tests were used to compare smoking rates and use of lipid-lowering drug.

On the basis of significant Hawaii-Japan univariate differences for nondietary and dietary variables in both men and women, sex-specific multiple linear regression models were used to examine the relations of dietary factors and physical activity to Hawaii-Japan differences in adiponectin concentration, with control for nondietary variables (21, 22). Because the distribution of serum adiponectin was positively skewed, a log-arithmic transformation was used to normalize the distribution. The basic model (model 1) included age and an indicator for site (Hawaii = 1, Japan = 0)toobtain the age-adjusted coefficient for Hawaii-Japan differences in log-transformed adiponectin concentrations. The addition of physical activity to model 1 was also examined. In addition to age, model 2 included BMI. Then, each dietary factor possibly explaining Hawaii-Japan differences in the adiponectin concentration was added to model 2 separately, and the percentage reduction in the site coefficient was calculated to assess influence of the added variable on Hawaii-Japan log-transformed adiponectin differences. Finally, dietary variables were included in combinations to assess their joint effect on Hawaii-Japan differences in log-transformed adiponectin concentrations. Sensitivity analyses were performed for the regression analyses by including participants with serum CRP concentrations > 10 mg/L, by excluding participants on any special diet, or by excluding participants with diabetes. SAS version 9.1.3 for WINDOWS (SAS Institute Inc, Cary, NC) was used throughout the analyses.

RESULTS

Demographic characteristics and blood chemistry concentrations

For both men and women, mean adiponectin concentration was significantly higher in Japan than in Hawaii (Table 1). Mean LDL cholesterol, glycated hemoglobin in men and women, and CRP in women were higher in Hawaii than in Japan. Average height was similar for Japanese in Japan and Japanese in Hawaii (Table 1). Average body weight and BMI were significantly higher in Hawaii than in Japan (P < 0.0001 for both) for both men and women with the difference being greater for men than for women. Smoking rate in men and mean for moderate + heavy physical activity (in h/d) in men and women were higher in Japan than in Hawaii (P < 0.0001). Antihypertensive and lipid-lowering drugs were used by a significantly greater percentage of persons of both sexes in Hawaii than in Japan.

TABLE 1.

Demographic characteristics and blood chemistry concentrations by sex of the INTERLIPID Study, Aito Town, Japan, and Honolulu, HI, 1997–1999¹

	Men			Women
Variable	Japan (n = 124)	Hawaii (n = 99)	P	Japan (n = 125)	Hawaii (n = 104)	P
Age (y)	49.4 ± 6.12	50.7 ± 5.1	0.10	49.8 ± 6.2	49.6 ± 4.9	0.79
Height (m)	1.67 ± 0.06	1.67 ± 0.06	0.89	1.54 ± 0.05	1.53 ± 0.05	0.55
Weight (kg)	65.4 ± 9.0	80.1 ± 14.8	<0.0001	55.3 ± 7.6	60.2 ± 11.7	0.0004
BMI (kg/m2)	23.3 ± 2.7	28.5 ± 4.6	<0.0001	23.2 ± 2.9	25.5 ± 5.0	0.0001
Smokers (%)	53.0	10.2	<0.0001	0.9	4.8	0.09
Moderate or heavy physical activity (h/d)	4.9 ± 4.7	2.2 ± 2.8	<0.0001	5.1 ± 4.5	1.1 ± 1.6	<0.0001
Hypertension (%)	22.2	32.7	0.09	13.1	24.8	0.03
Diabetes (%)	2.4	7.1	0.10	0.9	3.7	0.16
HTN Rx (%)	7.6	29.6	<0.0001	9.3	23.8	0.005
Lipid Rx (%)	4.4	18.4	0.002	2.8	10.8	0.02
Postprandial time (h) Serum concentrations	3.2 ± 2.2	4.2 ± 4.2	<0.0001	2.9 ± 2.0	3.6 ± 0.7	0.07
Adiponectin (μg/mL)	10.5 ± 5.5	6.7 ± 3.2	0.002	12.9 ± 5.9	9.4 ± 4.2	<0.0001
Total cholesterol (mg/dL)	203.8 ± 29.7	210.7 ± 28.5	0.09	205.4 ± 30.0	210.8 ± 32.6	0.21
HDL cholesterol (mg/dL)	53.4 ± 14.3	50.4 ± 10.2	0.07	61.5 ± 16.5	59.9 ± 13.0	0.43
LDL cholesterol (mg/dL)	125.8 ± 27.6	135.0 ± 27.7	0.02	125.8 ± 29.2	136.4 ± 33.6	0.015
CRP (mg/L)	0.84 ± 1.09	1.1 ± 1.3	0.09	0.58 ± 0.99	1.73 ± 1.93	<0.0001
Hb A1c (%)	4.7 ± 0.6	5.0 ± 0.8	0.002	4.5 ± 0.4	4.7 ± 0.7	0.0002

¹HTN Rx, on antihypertensive medication; lipid Rx, on lipid-lowering drug; CRP, C-reactive protein; Hb A1c, glycated hemoglobin.

²$\overline{x}$± SD (all such values).

Nutrient intakes

Most displayed nutrient intakes were significantly different across the 2 samples (Table 2). Mean intakes (%kcal) of non-vegetable protein, total fat, SFAs, monounsaturated fatty acids, PUFAs, n–6 PUFAs, and ARA in men and women were higher in Hawaii than in Japan. Participants in Japan consumed higher percentages of total calories from carbohydrates, vegetable protein, n–3 PUFAs, EPA, DHA, DPA, and more cholesterol than in Hawaii (P < 0.0001). Male participants in Japan consumed higher percentages of total calories from alcohol (P < 0.0001). Mean Keys dietary lipid score in men was significantly higher in Hawaii than in Japan (P < 0.0001).

TABLE 2.

Nutrient intakes and mean of four 24-h dietary recalls per person by sex of the INTERLIPID Study, Aito Town, Japan, and Honolulu, HI¹

	Men			Women
Variable	Japan (n = 124)	Hawaii (n = 99)	P	Japan (n = 125)	Hawaii (n = 104)	P
Total energy (kcal/d)	2390 ± 5352	2442 ± 609	0.50	1898 ± 379	1764 ± 391	0.01
Protein (%kcal)	15.3 ± 2.2	17.1 ± 2.8	<0.0001	16.3 ± 2.4	16.5 ± 2.9	0.53
Vegetable protein (%kal)	7.1 ± 1.0	6.0 ± 1.5	<0.0001	7.7 ± 1.2	6.1 ± 1.8	<0.0001
Nonvegetable protein (%kal)	8.2 ± 2.5	11.2 ± 3.1	<0.0001	8.7 ± 2.7	10.5 ± 3.4	<0.0001
Total available carbohydrate (%kcal)	56.4 ± 7.5	48.1 ± 7.3	<0.0001	57.8 ± 5.8	51.1 ± 7.6	<0.0001
Total fat (%kcal)	21.7 ± 4.7	31.9 ± 6.0	<0.0001	25.1 ± 4.3	31.7 ± 6.8	<0.0001
SFA (%kcal)	5.55 ± 1.49	9.16 ± 2.16	<0.0001	6.75 ± 1.69	9.36 ± 2.67	<0.0001
MUFA (%kcal)	7.72 ± 1.92	11.9 ± 2.6	<0.0001	8.92 ± 1.88	11.6 ± 2.8	<0.0001
PUFA (%kcal)	5.79 ± 1.45	7.56 ± 1.77	<0.0001	6.51 ± 1.18	7.45 ± 2.23	0.0002
n–3 PUFA (%kcal)	1.15 ± 0.33	0.89 ± 0.27	<0.0001	1.30 ± 0.36	0.87 ± 0.31	<0.0001
n–6 PUFA (%kcal)	4.62 ± 1.30	6.77 ± 1.66	<0.0001	5.18 ± 1.06	6.69 ± 2.06	<0.0001
ARA (%kcal)	0.06 ± 0.02	0.08 ± 0.03	<0.0001	0.07 ± 0.03	0.08 ± 0.04	0.02
EPA (%kcal)	0.14 ± 0.08	0.05 ± 0.06	<0.0001	0.15 ± 0.10	0.04 ± 0.05	<0.0001
DHA (%kcal)	0.25 ± 0.12	0.10 ± 0.10	<0.0001	0.27 ± 0.16	0.08 ± 0.10	<0.0001
DPA (%kcal)	0.04 ± 0.03	0.02 ± 0.02	<0.0001	0.04 ± 0.03	0.02 ± 0.02	<0.0001
EPA + DHA + DPA (%kcal)	0.43 ± 0.22	0.18 ± 0.17	<0.0001	0.47 ± 0.29	0.14 ± 0.16	<0.0001
α-Linolenic acid (%kcal)	0.75 ± 0.24	0.70 ± 0.21	0.81	0.80 ± 0.22	0.73 ± 0.29	0.04
Dietary cholesterol (mg/1000 kcal)	185.4 ± 65.9	128.3 ± 49.1	<0.0001	201.6 ± 74.9	132.4 ± 53.8	<0.0001
Total dietary fiber (g/1000 kcal)	6.92 ± 1.74	8.25 ± 2.29	<0.0001	9.80 ± 2.65	9.15 ± 3.47	0.13
Keys dietary lipid score	28.9 ± 6.4	31.2 ± 7.6	<0.0001	30.4 ± 6.7	32.1 ± 9.4	0.13
Alcohol (%kcal)	3.7 ± 5.5	2.8 ± 4.6	<0.0001	0.7 ± 1.7	0.6 ± 1.84	0.88

¹SFA, saturated fatty acid; MUFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid; ARA, arachidonic acid; EPA, eicosapentaenoic acid; DHA, docosahexaenoic acid; DPA, docosapentaenoic acid.

^{2$\overline{x}\pm \mathrm{S}\mathrm{D}$} (all such values).

Relation of dietary variables to Hawaii-Japan differences in adiponectin

Coefficients from multiple linear regression models used to examine relations of factors to Hawaii-Japan difference in log-transformed adiponectin concentration in men are shown in Table 3. The addition to model 1 of BMI reduced the coefficient for Hawaii-Japan log-transformed adiponectin difference by 92.3% (model 2). The P value for the Aito Town-Honolulu difference in serum adiponectin decreased from 0.0009 to 0.82. Further addition of dietary variables to model 2 did not change the coefficient significantly. Thus, in men, the sizable difference in BMI across the 2 populations statistically accounted for most of the difference in serum adiponectin.

TABLE 3.

Relation of variables to Hawaii-Japan log-transformed serum adiponectin difference in men in the INTERLIPID Study, Aito Town, Japan, and Honolulu, HI¹

Model	Coefficient for site (Japan = 0)	P for site	P for total R2	Coefficients for			Change in site coefficient from model 1
				Age	BMI	Nutrient
							%
Model 12	−0.08983	0.0009	0.0038	0.0007
+ Physical activity	−0.0753	0.008	0.0031	0.0004		0.005 (physical activity)	−16.7
Model 24	−0.0069	0.82	<0.0001	0.001	−0.0163		−92.3
+ Physical activity	0.0028	0.93	<0.0001	0.001	−0.0163	0.004 (physical activity)	−103.1
+ Nonvegetable protein	0.070	0.83	<0.0001	0.001	−0.0163	−0.005	−107.8
+ SFA	−0.039	0.31	<0.0001	0.002	−0.0173	0.010	−56.6
+ MUFA	−0.034	0.36	<0.0001	0.002	−0.0173	0.007	−61.8
+ Total n−3 PUFA	0.0007	0.98	<0.0001	0.001	−0.0173	0.021	−100.7
+ EPA + DHA + DPA	−0.005	0.90	<0.0001	0.001	−0.0163	0.008	−95.0
+ α-Linolenic acid	−0.006	0.85	<0.0001	0.001	−0.0163	0.029	−93.5
+ ARA	−0.005	0.87	<0.0001	0.001	−0.0163	−0.113	−94.1
+ Total fiber	−0.021	0.53	<0.0001	0.0009	−0.0153	0.008	−76.7
+ Dietary cholesterol	−0.013	0.70	<0.0001	0.001	−0.0163	−0.00001	−85.6
+ Keys score	−0.040	0.30	<0.0001	0.001	−0.0173	0.002	−86.7
+ Alcohol	−0.012	0.70	<0.0001	0.001	−0.0163	−0.001	−86.1

¹Coefficients for multiple linear regression models were used to examine the relations of dietary factors to Hawaii-Japan differences in log-transformed adiponectin concentration in men (124 in Japan and 99 in Hawaii), with control for nondietary variables as shown. P values for site coefficient and regression total R² are presented. Physical activity, moderate or heavy physical activity, usual number of hours per day; SFA, saturated fatty acid; MUFA, mono unsaturated fatty acid; PUFA, polyunsaturated fatty acid; EPA, eicosapentaenoic acid; DHA, docosahexaenoic acid; DPA, docosapentaenoic acid; ARA, arachidonic acid.

²Adjusted for site and age.

³P < 0.01 for the other coefficients.

⁴Adjusted for site, age, and BMI.

Coefficients from multiple linear regression models used to examine relations of factors to the Hawaii-Japan difference in log-transformed adiponectin concentration in women are shown in Table 4. The addition to model 1 of BMI reduced the coefficient for Hawaii-Japan adiponectin difference by 28.8% (model 2).

TABLE 4.

Relation of variables to Hawaii-Japan log-transformed serum adiponectin difference in women in the INTERLIPID Study, Aito Town, Japan, and Honolulu, HI¹

Model	Coefficient for site (Japan = 0)	P for site	P for total R2	Coefficients for			Change in site coefficient from model 1
				Age	BMI	Nutrient
Model 12	−0.13453	<0.0001	0.0038	0.0073			%
+ Physical activity	−0.1173	<0.0001	0.0031	0.00063		0.004 (physical activity)	−13.2
Model 24	−0.0963	0.0003	<0.0001	0.0073	−0.0173		−28.8
+ Physical activity	−0.0755	0.0003	<0.0001	0.0073	−0.0173	0.005 (physical activity)	−44.1
+ Nonvegetable protein	−0.0973	0.0004	<0.0001	0.0073	−0.0173	0.001	−27.7
+ SFA	−0.1333	<0.0001	<0.0001	0.0093	−0.0183	0.0155	−0.9
+ MUFA	−0.1173	0.0001	<0.0001	0.0083	−0.0183	0.008	−13.3
+ Total n−3 PUFA	−0.0993	0.0019	<0.0001	0.0073	−0.0173	−0.007	−26.3
+ EPA + DHA + DPA	−0.1333	0.0009	<0.0001	0.0083	−0.0173	−0.108	−1.0
+ α-Linolenic acid	−0.0903	0.0008	<0.0001	0.0073	−0.0173	0.079	−33.1
+ ARA	−0.0943	0.0005	<0.0001	0.0073	−0.0173	−0.153	−29.8
+ Total fiber	−0.0973	0.0003	<0.0001	0.0009	−0.0153	0.008	−27.5
+ Dietary cholesterol	−0.0873	0.0046	<0.0001	0.0073	−0.0183	0.0001	−35.1
+ Keys score	−0.0993	0.0002	<0.0001	0.0083	−0.0183	0.00355	−26.3
+ Alcohol	−0.0963	0.0003	<0.0001	0.0073	−0.0173	0.008	−28.6
+ Physical activity + nonvegetable	−0.59	0.074	<0.0001	0.0073	−0.0183	0.005 (physical activity), 0.003	−56.1
protein + total n−3 PUFA +						(nonvegetable protein),
ARA + cholesterol						−0.013 (total n−3 PUFA), −0.69 (ARA), 0.0003 (cholesterol)

¹Coefficients for multiple linear regression models were used to examine the relations of dietary factors to Hawaii-Japan differences in log-transformed adiponectin concentration in women (125 in Japan and 104 in Hawaii), with control for nondietary variables as shown. P values for site coefficient and regression total R² are presented. Physical activity, moderate or heavy physical activity, usual number of hours per day; SFA, saturated fatty acid; MUFA, monounsaturated fatty acid; PUFA, polyunsaturated fatty acid;, EPA, eicosapentaenoic acid; DHA, docosahexaenoic acid; DPA, docosapentaenoic acid; ARA, arachidonic acid.

²Adjusted for site and age.

³P < 0.01 for the other coefficients.

⁴Adjusted for site, age, and BMI.

⁵P < 0.05 for the other coefficients.

With the addition of physical activity to model 2, percentage of reduction in the coefficient increased to 44.1%, but contribution of physical activity to the regression was not statistically significant (P = 0.17), and coefficient for log-transformed adiponectin difference remained significant. Although the addition to model 2 of SFA and Keys dietary lipid score, singly, had a significant independent relation to adiponectin (both P < 0.01), these dietary variables singly did not reduce the coefficient for log-transformed adiponectin difference significantly. With the addition to model 2 of nonvegetable protein, n–3 PUFA, ARA, and dietary cholesterol intakes together, the percentage of reduction in the coefficient increased to 56.1%. The P value for the Aito Town-Honolulu difference in log-transformed serum adiponectin decreased from <0.0001 to 0.074. However, none of these nutrient variables, considered singly or in combination, had a significant independent relation to adiponectin. Because the effect of BMI and physical activity on the adiponectin difference in women was much smaller than in men, we made interaction analyses between sex and BMI and between sex and physical activity on the adiponectin difference for men and women combined data. However, neither analyses show significant interactions (P = 0.22, P = 0.77, respectively). Similar results were observed after including those with serum CRP concentrations > 10 mg/L, excluding the participants on any therapeutic diet, or excluding the participants with diabetes.

DISCUSSION

The main findings here were serum adiponectin concentrations were significantly higher in population samples of Japanese in Japan than Japanese Americans in Hawaii, people of the same genetic background. In men, adjustment for lower Japanese BMI reduced the log-adiponectin Japan-Hawaii difference by 92%, to a nonsignificant level, whereas in women, BMI reduced this difference by only 29%. Further adjustment for 4 nutrients in women produced a further reduction in the difference, although none of the nutrient variables, considered singly or in combination, had a significant independent relation to adiponectin.

There have been several studies on relation of body weight to adiponectin concentration. Esposito et al (23) found in a randomized controlled trial that a Mediterranean-style diet and increased physical activity for 2 y significantly decreased BMI and concomitantly increased adiponectin concentration in postmenopausal obese women. Arawaka et al (24) in a 5-y observational study involving > 2000 participants found that change in body weight inversely correlated well with change in adiponectin concentration. In contrast, several short-term weight reduction intervention studies failed to show significant change in adiponectin concentration, despite significant reduction in body weight (25–29). Liu et al (27) observed a significant rise in adiponectin mRNA expression in subcutaneous adipose tissue after a 2-d low-caloric diet in morbidly obese women, whereas circulating adiponectin concentration remained unchanged. Thus, it appears that adiponectin is involved mainly in long-term rather than short-term regulatory mechanisms (18). In addition, it was reported that there is only limited within-person variation in the plasma adiponectin concentrations during a period of 1 y with control for changes in BMI, and it did not change postprandially (17, 18).

There have been few studies on the association between habitual dietary intakes and adiponectin concentrations. A cross-sectional study in 114 Greek students by Yannakoulia et al (30) showed no correlation between adiponectin concentrations and total energy intake, protein intake (expressed as a percentage of energy intake), or fat or carbohydrate intake. However, in 780 diabetic men from the Health Professionals Follow-up Study, Qi et al (31) found in their cross-sectional study that diets low in glycemic load and high in fiber to be associated with higher adiponectin concentrations.

When we found that the mean adiponectin concentration was significantly higher in Japan than in Hawaii, our initial expectation was that long-chain n–3 PUFAs from fish would be nutrients related to the difference because their consumption in Japan was strikingly higher (≈3 times) than in Hawaii. In addition, a previous experimental study in animals showed that fish oil feeding raised adiponectin concentrations (32). However, in multiple regression analyses adjusted for age and site, EPA, DHA, and DPA in combination did not reduce the coefficient for the serum adiponectin difference. Furthermore, what we regard as generally adverse dietary factors, such as higher SFA and Keys dietary lipid score, appeared to raise adiponectin in women, and nutrients generally regarded as favorable, such as EPA, DHA, and DPA in combination, did not influence the adiponectin difference.

In our previous comparative study, American men with higher BMI had substantially higher adiponectin concentrations than did Japanese men (33, 34). In the present study, significantly higher mean serum adiponectin concentration was found in Japan than in Hawaii, statistically attributable to differences in BMI, especially in men; it was statistically attributable in women largely to differences in BMI and physical activity.

The main strengths of the present study are 1) use of population-based samples, 2) standardized collection of high-quality blood pressure and nutrition data, 3) use of an improved nutrient database, and 4) use of multiple quality-control procedures. The study was limited by its 2-sample cross-sectional design and small number of participants. Findings may or may not be generalizable to all Japanese and to other populations. Furthermore, despite high-quality acquisition methods of dietary nutrient data, limitations in the reliability may still bias against the nutritional variables weighing heavily in the multivariate models.

In conclusion, the mean adiponectin concentration was significantly higher in Japan than in Hawaii for both sexes. In men, this difference was related overwhelmingly to BMI; in women, it was related to BMI and physical activity, with 4 specific nutrients playing a modest statistically nonsignificant role in further reduction of the adiponectin difference.