Low serum testosterone level as a predictor of increased visceral fat in Japanese-American men


OBJECTIVE: To examine the association between baseline testosterone levels and changes in visceral adiposity in Japanese-American men.

DESIGN: Prospective observational study.

SUBJECTS: Second-generation Japanese-American males enrolled in a community-based population study.

MEASUREMENTS: At baseline, 110 men received a 75 g oral glucose tolerance test (OGTT), and an assessment of body mass index (BMI); visceral adiposity measured as intra-abdominal fat area (IAF) using computed tomography (CT); fasting insulin and C-peptide levels; and total testosterone levels. IAF was re-measured after 7.5 y. Subcutaneous fat areas were also measured by CT in the abdomen, thorax and thigh. The total fat (TF) was calculated as the sum of IAF and total subcutanous fat areas (SCF).

RESULTS: After 7.5 y, IAF increased by a mean of 8.0 cm2 (95% CI: 0.8, 15.3). Baseline total testosterone was significantly correlated with change in IAF (r=−0.26, P=0.006), but not to any appreciable degree with change in BMI, TF, or SCF. In a linear regression model with change in IAF as the dependent variable, baseline testosterone was significantly related to this outcome while adjusting for baseline IAF, SCF, BMI, age, diabetes mellitus status (OGTT by the WHO diagnostic criteria) and fasting C-peptide (regression coefficient for baseline testosterone [nmol/l]=−107.13, P=0.003).

CONCLUSIONS: In this Japanese-American male cohort, lower baseline total testosterone independently predicts an increase in IAF. This would suggest that by predisposing to an increase in visceral adiposity, low levels of testosterone may increase the risk of type 2 diabetes mellitus.


Obesity is a major risk factor in the development of type 2 diabetes millitus (DM), hypertension and cardiovascular disease (CVD).1,2,3 In particular, central obesity with a preponderance of intra-abdominal fat (IAF) accumulation correlates more strongly with the estimates of disease risk,4,5 and is known to be more diabetogenic6,7 than subcutaneous fat accumulation. Visceral fat or IAF area also contributes significantly more to the variances in cardiovascular risk profile than waist-to-hip ratio (WHR) measurement,5 which does not differentiate between subcutaneous- and intra-abdominal fat. Assessing body fat distribution by computed tomography (CT), which gives accurate estimates of IAF area,8 would permit examination of the contribution of visceral fat to glucose homeostasis and lipid metabolism. In fact, CT-measured IAF areas were found to be significantly increased at baseline in Japanese-American subjects who later developed type 2 DM, compared with those who remained non-diabetic.9

Men are at higher risk than women of developing coronary heart disease at an earlier age and tend to gain weight predominantly in the abdomen. Hyperandrogenicity is further implicated in the regulation of body fat deposition in women with central obesity and insulin resistance.10,11 However, in contrast to the findings in women, several studies have found that greater central obesity (based on WHR) is actually associated with lower levels of androgens in men.12,13,14,15,16,17,18 Numerous studies have also clearly documented a cross-sectional association between lower testosterone concentrations in men and higher circulating insulin and insulin resistance.14,15,19,20,21,22,23 In fact, an earlier study showed that low-dose testosterone administration in obese men decreased CT-measured visceral fat mass without changing subcutaneous fat or lean body mass and also reduced insulin resistance.24 Diabetic men have also been found to have relatively low testosterone levels.25

Björntorp has proposed that relative hypogonadism might be the primary event preceding the development of central obesity and insulin resistance in men. Low testosterone levels have been shown to predict the risk of developing diabetes in Multiple Risk Factor Intervention Trial (MRFIT) men16 and in older Swedish men,17 but this association was independent of WHR in the latter population. CT assessment of abdominal fat was not performed in either study. The question remains in men as to which abdominal fat compartment (visceral or intra-abdominal vs subcutaneous) is principally associated with testosterone, and whether low testosterone level precedes or follows gain in visceral adiposity.

We therefore tested the hypothesis that low baseline testosterone levels were associated with IAF accumulation in Japanese-American men followed prospectively for 7.5 y in King County, Washington State. The relationship between testosterone levels and indices of glucose metabolism at baseline were also examined since Japanese Americans are known to have high prevalence of type 2DM.26


Study subjects

The study population comprised second-generation Japanese-American (Nisei) men enrolled in the Japanese-American Community Diabetes Study (JACDS). Details about the selection and recruitment of the sample population have been described previously.27 Briefly, subjects were chosen from among volunteers and were representative of the Japanese Americans in King County, WA, in age distribution, residence and parental immigration patterns. The enrollment of 229 Nisei men started in 1983 and all follow-ups were completed in 1994. Serum total testosterone was first measured at 2.5 y of follow-up (‘baseline'’ for this analysis) when those men who were not diabetic at study entry returned for an oral glucose tolerance test (OGTT), CT measurement of body fat distribution, and measurements of serum testosterone, plasma insulin and C-peptide. All measurements except testosterone were repeated 7.5 y later. Of the original cohort of 229 Nisei men, 151 were not diabetic at entry, 135 had baseline CT measurement of IAF at 2.5 y, and 113 returned for the 7.5 y follow-up and repeat IAF assessment. Of these, 110 subjects had a baseline total testosterone measurement and constituted the study sample for this analysis.

Data collection

Baseline and follow-up evaluations were performed at the General Clinical Research Center at the University of Washington Medical Center. The research protocol was reviewed and approved by the institutional human subjects review committee and signed informed consent was obtained from all participants. A 75 g OGTT was used to classify all subjects as having normal glucose tolerance (NGT), impaired glucose tolerance (IGT), or type 2 DM, based on the WHO diagnostic criteria.28 Serum glucose was assayed by an automated glucose oxidase method at 0, 30, 60, 90, 120 and 180 min during the OGTT. Plasma insulin and C-peptide levels were determined by radio immunoassay performed at the Diabetes Endocrinology Research Center at times 0, 30, 60, 90, 120, and 180 min during the OGTT. Serum total testosterone level was measured in the Department of Laboratory Medicine at the University of Washington Medical Center using a commercially prepared radio immunoassay kit (Immuchem Corp, Carson, CA). The sensitivity of the assay was 0.69 and 51.9 nmol/1 for the lower and upper limits, respectively. The reference range was 9.69–35 nmol/1 in adult males and the interassay coefficient of variation was between 7.5–8.8%

Mesurement of body mass index and body fat distribution

Body mass index (BMI) was computed as weight in kilograms divided by height in meters squared (kg/m2). Single computed tomography (CT) scans of 10 mm slice thickness were obtained of the thorax, abdomen and right thigh, as previously described.6 Adipose tissue was within the attenuation range of −250 and −50 Hounsfield Units (HU). Visceral obesity was measured as intra-abdominal fat (IAF) area in centimeters squared (cm2) at the level of the umbilicus (approximately L4–5). Subcutaneous fat areas were also measured by CT scan in the abdomen, thorax and thigh. Total fat (TF) was calculated as the sum of IAF and total subcutaneous fat (SCF) areas.

Statistical analysis

Base 10 log transformed testosterone levels were calculated to correct for the slight positive kurtosis of the non-transformed total testosterone values. The log-transformed testosterone levels were used in all other statistical computations. Means of differences between baseline and 7.5 y follow-up were compared using paired t-test or Wilcoxon paired signed-ranks test in cases where the assumption of normality was not satisfied. Pearson correlation coefficients were calculated for univariate comparisons of continuous variables, except in cases of skewed distributions, when Spearman rank coefficients were calculated. Statistical significance of a linear trend was computed using ANOVA. Multiple linear regression analysis was used to model the change in IAF as a function of variables of interest and the presence of interaction between covariates in the models. Analysis of residuals was performed to examine model fit and adherence to regression assumptions.


Baseline characteristics of study subjects

Table 1 shows the baseline characteristics (mean±s.d.) of the 110 men as a group and also by their baseline OGTT status. The IGT and DM groups had significantly higher levels of fasting insulin and C-peptide. Moreover, baseline BMI, TF, IAF and SCF were also significantly greater in the IGT and DM groups. Baseline testosterone levels were lower in the DM group, at a borderline level of significance.

Table 1 Baseline characteristics of subjects, total and by OGTT status

Changes in BMI and fat distribution over 7.5 y

Over the follow-up period of 7.5 y, the BMI of most subjects (68%) increased. This was also reflected by a gain in TF as well. Although the quantities of fat areas (cm2) increased to a similar degree for both IAF and SCF (see Table 2), the proportional increase was greater for IAF (6.7%) than for SCF (2.6%). Statistically, IAF also increased to a significant degree during follow-up, while SCF did not (Table 2).

Table 2 Changes in BMI and fat distribution over 7.5 y

Correlation of testosterone with changes in BMI and fat distribution

Change in IAF over 7.5 y (ΔIAF) paralleled change in BMI (ΔBMI; r=0.53, P<0.0001), change in TF (ΔTF; r=0.73, P<0.0001), and change in SCF (ΔSCF; r=0.36, P<0.0001). However, only ΔIAF correlated significantly with baseline testosterone (r=−0.26, P=0.006; Figure 1). No significant association was detected for baseline testosterone vs ΔBMI (r=−0.08), ΔTF (r=−0.10), or ΔSCF (r=−0.04).

Figure 1

Correlation between intra-abdominal fat change and baseline testosterone.

Categorical trend

When ΔIAF was grouped into quartiles (see Table 3), a significant linear trend (P=0.029) was detected across the means of the baseline testosterone levels, adjusted for baseline BMI, IAF, SCF, age and OGTT status. The fourth quartile, which had the greatest gain in mean intra-abdominal fat area, had the lowest log-transformed testosterone level (1.17 nmol/1). Conversely, the subjects in the first quartile, all of whom had lost intra-abdominal fat over the 7.5 y, had a relatively high mean testosterone level (1.23 nmol/1) at baseline. On the other hand, no significant linear trends were detected in baseline testosterone levels by quartiles of ΔTF, ΔSCF or ΔBMI (data not shown).

Table 3 Baseline testosterone levels by quartiles of intra-abdominal fat changes over 7.5 y

Correlation between baseline testosterone and insulin and C-peptide

Low baseline testosterone level correlated significantly with baseline levels of fasting C-peptide (r=−0.34, P<0.0001) and fasting insulin (r=−0.26, P<0.01). The strength of the association between fasting insulin and testosterone was diminished (P>0.05) after adjusting for baseline IAF, while the C-peptide level remained significantly correlated (P=0.001) with testosterone.

Multivariate linear regression between IAF change over 7.5 y and baseline testosterone

Linear regression was used to examine the relationship between ΔIAF and baseline testosterone. A multivariate model was used with adjustment for potential confounders, including baseline IAF and SCF areas, BMI, age, OGTT status (NGT, IGT and DM), and fasting C-peptide level. The main variable of interest, baseline testosterone, was a significant predictor of ΔIAF. The negative sign of its regression coefficient indicates an inverse association between baseline testosterone level and change in IAF. Additionally, no significant interactions were found among any pair of covariates. Table 4 shows the coefficients of the independent variables and the corresponding significance levels. Baseline SCF, as a potential confounder, was also associated with change in IAF area over 7.5 y (P=0.046).

Table 4 Multiple linear regression analysis of intra-abdominal fat change on baseline log-transformed testosterone with adjustment for baseline intra-abdominal fat, sub-cutaneous fat, BMI, age and OGTT status


This study is the first, to our knowledge, to demonstrate a strong inverse relationship between the baseline level of total testosterone and accumulation of CT-measured IAF area over time. Most other observational studies have reported the relationship between testosterone and central obesity using cross-sectional data or WHR, instead of CT quantified visceral fat area. Our results suggest that baseline serum testosterone level is related to the degree of IAF accumulation over 7.5 y of follow-up in a group of Japanese-American men. Furthermore, if we compared two hypothetical groups of men with the same baseline age, BMI, IAF and SCF areas, DM status and fasting C-peptide level, but with one group having total testosterone level at the upper limit of the ‘normal’ range (about 35 nmol/1) and the other group at the lower limit (about 10 nmol/1), the group with lower baseline testosterone level would gain an additional 60 cm2 of IAF in 7.5 y (based on regression coefficients in Table 4). If those observations are correct, then even a ‘normal’ plasma testosterone level at the lower end may predispose to greater visceral adiposity and consequently, higher risk for diabetes mellitus and coronary heart disease.

The distribution pattern of adipose tissue varies by gender. The android pattern, seen more often in men, is characterized by fat deposition in the abdomen, whereas femoral–gluteal fat accumulation, seen more often in women, characterizes the gynoid pattern. The common belief, therefore, is that maleness with excess androgens is associated with central obesity. While hyperandrogenicity is a significant contributor to increased central adiposity in women,11,29,30 it is clearly not so in men. Instead, testosterone levels are lower in men with increased abdominal girth or WHR compared to age-matched, non-obese men.12,13,31,32,33 To better quantify the contribution of IAF to this observation, a method such as CT is required, because of the imperfection correlation between surface circumference measurements and visceral fat mass.34

IAF depots are more metabolically active with more rapid lipid turnover35 and greater lipid uptake36 than subcutaneous fat depots. Omental adipocytes have been shown to be more lipolytic in response to catecholamines37,38 compared to subcutaneous adipocytes, possibly due to differences in he number and activity of β and α adrenoceptors in the two adipose tissues.39 Testosterone is shown to promote lipolysis by enhancing catecholamine-mediated lipid mobilization in rat adipocytes.40 Visceral fat tissue may be more sensitive to the testosterone effect seen in vitro. In a randomized study, a modest dose of testosterone replacement to middle-aged centrally obese men reduced CT-measured visceral fat after 8 months, whereas no significant change in subcutaneous fat was detected.24 Additionally, testosterone administration resulted in a greater inhibition of lipoprotein lipase (LPL), the main enzyme regulating lipid uptake and accumulation into adipocytes, in the abdominal subcutaneous fat than in the femoral fat.41,42 Although it has not been conclusively demonstrated, it has been inferred that testosterone inhibits LPL to a similar or greater degree in visceral fat as in abdominal subcutaneous fat. In another study, a one-time high-dose testosterone administration to apparently healthy men decreased lipid uptake in intra-abdominal fat but not in abdominal subcutaneous fat, suggesting that exogenous testosterone may inhibit LPL activity to a greater degree in intra-abdominal fat than abdominal subcutaneous fat.36 Higher density of androgen receptors present in intra-abdominal fat43 could partly explain why this fat depot is more responsive to testosterone replacement than subcutaneous fat and why there is IAF preferentially accumulated with lower serum testosterone.

Visceral obesity is strongly associated with the metabolic consequences of insulin resistance, such as impaired glucose metabolism and unfavorable lipid profile. This association is also clearly demonstrated in our Japanese-American cohorts from earlier studies.44,45 It has been postulated that increased plasma FFA levels are associated with insulin resistance.46,47 By releasing directly into the portal system a greater amount of FFA from enhanced triglyceride breakdown, as a result of greater catecholamine-stimulated and/or reduced insulin-inhibited lipolysis, visceral fat may contribute more profoundly to the metabolic abnormalities seen with the insulin resistance syndrome.

Many studies have shown an association between relative hypogonadism and insulin resistance.14,15,19,20,21,22,23 We also have previously shown that increased fasting insulin and C-peptide concentrations predicted IAF accumulation in Japanese Americans.48 It is, therefore, not surprising that in the current analysis, both fasting insulin and C-peptide, as markers of insulin resistance and pancreatic β-cell function, respectively, were inversely correlated with total testosterone. However, the correlation between low testosterone and fasting insulin levels was diminished after adjusting for baseline IAF area. One possible explanation could be that IAF may be part of the pathway through which lower testosterone is related to insulin resistance. A similar observation was reported from 79 younger men in Quebec, Canada, in whom the authors found the close cross-sectional association between plasma testosterone and insulin levels (r=−0.25) to be largely mediated by the concomitant variation in CT-measured visceral abdominal fat.49 On the other hand, in another cross-sectional study of 23 healthy young Swedish men, Seidell et al observed that free testosterone levels remained independently correlated with fasting insulin (r=−0.66 without adjustment), even after adjusting for CT-visceral fat area.14 Haffner also reported a significant correlation between free testosterone level and fasting insulin after adjusting for BMI and WHR.23 Visceral adiposity was not assessed by CT in that study. No conclusion about the possible temporal relationship among testosterone, IAF and insulin resistance can be reached from these cross-sectional analyses. Overall, our results from the prospective analysis are in agreement with the temporal sequence of events proposed by Björntorp: relative hypogonadism precedes central obesity, a notion that is supported by the in vivo effect of low-dose testosterone administration in reducing visceral fat and improving insulin sensitivity.24 Additionally, our cross-sectional analysis also suggests, but does not confirm, that testosterone may play a role in the insulin resistance syndrome by modulating the degree of visceral adiposity.

Since metabolism of androgens is closely linked to the protein-binding state of testosterone, several studies have also examined the importance of free testosterone and SHBG on obesity and DM prevalence. It appears that in moderately obese men (BMI<35), only total testosterone concentration was decreased, primarily as a result of reduced SHBG-binding capacity, and there was no reduction in the unbound or free testosterone fraction or any impairment in the pituitary–gonadal axis.13 In massively obese men (BMI>40), however, the LH pulse amplitude was reduced, resulting in reduction in both total and free testosterone levels.13 Therefore, the inconsistent findings in the literature on free testosterone levels in obese men 14,21,22,23 may have been due to admixture of both modestly and massively obese subjects. Moreover, free testosterone demonstrated weaker associations than total testosterone with indices of glucose metabolism in middle-aged healthy Finns.15 Overall evidence, thus, would suggest that total testosterone is a better marker in examining relationships of sex steroids, obesity and glucose metabolism in men. Free testosterone levels were not assessed in our subjects due to the unavailability of a reliable assay at the time of data collection. However, since the mean BMI in our study population was 25.3 kg/m2, their free testosterone concentrations were most likely in the normal range.

We did not measure SHBG concentrations and were not able to directly examine the relative roles of SHBG and bio-available testosterone levels in IAF accumulation or glucose metabolism. SHBG level, as an indirect index of bio-available fraction of sex steroids, has been proposed to be a better marker for androgen metabolism and insulin resistance. Similar to testosterone, decreased SHBG concentrations were also associated with higher CT-measured visceral fat areas in Swedish men14 and with indices of glucose metabolism and WHR in Finns.15 In the MRFIT trial, both low SHBG and testosterone levels predicted the development of type 2 DM in men.16 However, in older Swedish men, low testosterone was a more significant predictor of the onset of diabetes than low SHBG level,17 and in Mexican-American males, SHBG level was not a predictor of type 2 DM at all.50 Thus, SHBG does not clearly appear to be superior to testosterone level with regard to strength of association with measures of abnormal glucose metabolism.

The limitations of our study are several. Firstly, these observations in Japanese-American men may not apply to other ethnic populations. Nevertheless, our findings are consistent with those found in Swedes, showing an association of testosterone and central obesity. Another limitation was that SHBG was not measured in our subjects. However, as discussed above, this is unlikely to have decreased our power to detect an association between androgenicity and central obesity. Finally, insulin sensitivity was assessed indirectly by fasting insulin level. As a marker, fasting insulin level correlated fairly well with insulin resistance quantified by euglycemic hyperinsulinemic clamp, and is, therefore, a reasonable surrogate measure of insulin sensitivity in population studies.51 Further studies are needed to confirm our findings in other populations. If confirmed, it may be appropriate to examine the long-term effect of exogenous testosterone upon IAF in individuals with low normal testosterone levels.


  1. 1

    Ohlson LO, Larsson B, Svärdsudd K, Welin L, Eriksson H, Wilhelmsen L. Björntorp P, Tibblin G . The influence of body fat distribution on the incidence of diabetes mellitus Diabetes 1985 34: 1055–1058.

    CAS  Article  Google Scholar 

  2. 2

    Bottiger LE, Carson LA . Risk factors for ischemic vascular death for men in the Stockholm Prospective Study Atherosclerosis 1980 36: 389–408.

    Article  Google Scholar 

  3. 3

    Hubert HB, Feinleib M, McNamara PM, Castelli WP . Obesity as an independent risk factor for cardiovascular disease: a 26-year follow-up of participants in the Framingham Heart Study Circulation 1983 67: 968–977.

    CAS  Article  Google Scholar 

  4. 4

    Segal KR, Dunaif A, Gutin B, Albu J, Nyman A, Pi-Sunyer FX . Body composition, not body weight, is related to cardiovascular disease risk factors and sex hormone levels in men. J Clin Invest 1987 80: 1050–1055.

    CAS  Article  Google Scholar 

  5. 5

    Peiris AN, Sothman MS, Hoffman RG, Hennes MI, Wilson CR, Gustafson AB, Kissebah AH . Adiposity, fat distribution, and cardiovascular risk Ann Intern Med 1989 110: 867–872.

    CAS  Article  Google Scholar 

  6. 6

    Shuman WP, Morris LL, Leonetti DL, Wahl PW, Moceri VM, Moss AA, Fujimoto WY . Abnormal body fat distribution detected by computed tomography in diabetic men Invest Radiol 1986 21: 483–487.

    CAS  Article  Google Scholar 

  7. 7

    Sparrow D, Borkan GA, Gerzof SG, Wisniewski C, Silbert CK . Relationship of fat distribution to glucose tolerance. Results of computed tomography in male participants of the Normative Aging Study Diabetes 1986 35: 411–415.

    CAS  Article  Google Scholar 

  8. 8

    Thaete FL, Colberg SR, Burke T, Kelley DE . Reproducibility of computed tomography measurement of visceral adipose tissue area Int J Obes 1995 19: 464–467.

    CAS  Google Scholar 

  9. 9

    Bergstrom RW, Newell-Morris L, Leonetti DL, Shuman WP, Wahl PW, Fujimoto WY . Association of elevated fasting C-peptide and increased intra-abdominal fat distribution with development of NIDDM in Japanese-American men Diabetes 1990 39: 104–111.

    CAS  Article  Google Scholar 

  10. 10

    Rebuffé-Scrive M, Cullberg G, Lundberg P, Lindstedt G, Björntorp P . Anthropometric variables and metabolism in polycystic ovarian disease Horm Metab Res 1989 21: 391–397.

    Article  Google Scholar 

  11. 11

    Evans DJ, Hoffman RG, Kalkhoff RK, Kissebah AH . Relationship of androgenic activity to body fat topography, fat cell morphology, and metabolic aberrations in premenopausal women J Clin Endocrinol Metab 1983 57: 304–310.

    CAS  Article  Google Scholar 

  12. 12

    Glass AR, Swerdloff RS, Bray GA, Dahms W, Atkinson RL . Low serum testosterone and sex hormone binding globulin in massively obese men J Clin Endocrinol Metab 1977 45: 1211–1219.

    CAS  Article  Google Scholar 

  13. 13

    Giagulli VA, Kaufman JM, Vermeulen A . Pathogenesis of the decreased androgen levels in obese men J Clin Endocrinol Metab 1994 79: 997–1000.

    CAS  PubMed  Google Scholar 

  14. 14

    Seidell JC, Björntorp P, Sjöström LS, Kvist H, Sannerstedt R . Visceral fat accumulation in men is positively associated with insulin, glucose, and C-peptide levels, but negatively with testosterone levels Metabolism 1990 39: 897–901.

    CAS  Article  Google Scholar 

  15. 15

    Haffner SM, Karhapää P, Mykkänen L, Laakso M . Insulin resistance, body fat distribution and sex hormones in men Diabetes 1994 43: 212–219.

    CAS  Article  Google Scholar 

  16. 16

    Haffner SM, Shten J, Sern MP, Smith GD, Kuller L . Low levels of sex hormone-binding globulin and testosterone predict the development of non-insulin-dependent diabetes mellitus in men Am J Epidemiol 1996 143: 889–897.

    CAS  Article  Google Scholar 

  17. 17

    Tibblin G, Adlerberth A, Lindstedt G, Björntorp P . The pituitary–gonadal and health in elderly men Diabetes 1996 45: 1605–1609.

    CAS  Article  Google Scholar 

  18. 18

    Khaw KT, Barrett-Connor E . Lower endogenous androgens predict central adiposity in men Ann Epidemiol 1992 2:: 675–682.

    Article  Google Scholar 

  19. 19

    Phillips GB . Relationship between serum sex hormones and glucose, insulin, and lipid abnormalities in men with myocardial infarction Proc Nat Acad Sci USA 1979 74: 1729–1732.

    Article  Google Scholar 

  20. 20

    Simon D, Preziosi P, Barrett-Connor E, Roger M, Saint-Paul M, Nahoul K, Papoz L . Interrelation between plasma testosterone and plasma insulin in healthy adult men: the Telecom Study Diabetologia 1992 35: 173–177.

    CAS  Article  Google Scholar 

  21. 21

    Phillips GB . Relationship between sex hormones and the glucose insulin–lipid defect in men with obesity Metabolism 1993 42: 116–120.

    CAS  Article  Google Scholar 

  22. 22

    Pasquali R, Casimirri F, Cantobelli S, Melchionda N, Morselli-Labate AM, Fabbri R, Capelli M, Bortoluzzi L . Effect of obesity and body fat distribution on sex hormones and insulin in men Metabolism 1991 40: 101–104.

    CAS  Article  Google Scholar 

  23. 23

    Haffner SM, Valdez RA, Mykkänen L, Stern MP, Katz MS . Decreased free testosterone and dehydroepiandrosterone sulfate are associated with decreased glucose and insulin concentrations in non-diabetic men Metabolism 1994 43: 599–603.

    CAS  Article  Google Scholar 

  24. 24

    Mårin P, Holmäng S, Jönsson L, Sjöström L, Kvist H, Holm G, Lindstedt G, Björntorp P . The effect of testosterone treatment on body composition and metabolism in middle aged obese men Int J Obes 1992 16: 991–997.

    Google Scholar 

  25. 25

    Andersson B, Mårin P, Lissner L, Vermeulen A, Björntorp P . Testosterone concentrations in women and men with NIDDM Diabetes Care 1994 17: 405–411.

    CAS  Article  Google Scholar 

  26. 26

    Fujimoto WY, Bergstrom RW, Boyko EJ, Kinyoun JL, Leoneti DL, Newell-Morris LL, Robinson LR, Shuman WP, Stolov WC, Tsunehara CH, Wahl PW . Diabetes and diabetes risk factors in second- and third-generation Japanese Americans in Seattle, Washington Diabetes Res Clin Pract 1994 24: S43–S52.

    Article  Google Scholar 

  27. 27

    Fujimoto WY, Leonetti KL, Kinyoun JL, Newell-Morris L, Shuman WP, Stolov WC, Wahl PW . Prevalence of diabetes mellitus and impaired glucose tolerance among second generation Japanese-American men Diabetes 1987 36: 721–729.

    CAS  Article  Google Scholar 

  28. 28

    World Health Organization Expert Committee . Second report on diabetes mellitus, 1980 WHO Technical Report Series No. 646 1980.

    Google Scholar 

  29. 29

    De Pergola G, Giagulli Va, Garruti G, Cospite MR, Giorgino F, Cignarelli M, Giorgino R . Low dehydroepiandrosterone circulating levels in premenopausal obese women with very high body mass index Metabolism 1991 40: 187–190.

    CAS  Article  Google Scholar 

  30. 30

    Giagulli VA, De Pergola G, Giorgino F, Cignarelli M, Abbaticchio G, Vermeulen A, Giorgino R . Increased free testosterone but normal 5a-reduced testosterone metabolites in obese premenopausal women Clin Endocrinol Oxf 1992 36: 553–558.

    CAS  Article  Google Scholar 

  31. 31

    Stanik S, Dornfeld LP, Maxwell MH, Viosca SP, Korenman SG . The effect of weight loss on reproductive hormones. J Clin Endocrinol Metab 1981 53: 828–831.

    CAS  Article  Google Scholar 

  32. 32

    Glass AR, Burman KD, Dahms WT, Boehm TM . Endocrine function in obesity Metabolism 1981 30: 89–104.

    CAS  Article  Google Scholar 

  33. 33

    Vermeulen A, Kaufman JM, Deslypere JP, Thomas G . Attenuated LH pulse amplitude but normal LH pulse frequency and its relation to plasma androgens in hypogonadism of obese men J Clin Endocrinol Metab 1993 76: 1140–1146.

    CAS  PubMed  Google Scholar 

  34. 34

    Seidell JC, Björntorp P, Sjöström L, Sannerstedt R, Krotkiewski M, Kvist H . Regional distribution of muscle and fat mass in men—new insight into the risk of abdominal obesity using computed tomography Int J Obes 1989 13: 289–303.

    CAS  PubMed  Google Scholar 

  35. 35

    Mårin P, Andersson B, Ottosson M, Olbe L, Chowdhury B, Kvist H, Holm G, Sjöström L, Björntorp P . The morphology and metabolism and intra-abdominal adipose tissue in men Metabolism 1992 41: 1242–1248.

    Article  Google Scholar 

  36. 36

    Mårin P, Lönn L, Andersson B, Odén B, Olbe L, Bengtsson, Björntorp P . Assimilation of triglycerides in subcutaneous and intra-abdominal adipose tissue in vivo in men: effects of testosterone J Clin Endocrinol Metab 1996 81: 1018–1022.

    PubMed  Google Scholar 

  37. 37

    Hellmer J, Marcus C, Sonnenfeld T, Arner P . Mechanisms for differences in lipolysis between human subcutaneous and omental fat cell J Clin Endocrinol Metab 1992 75: 15–20.

    CAS  PubMed  Google Scholar 

  38. 38

    Richelsen B, Pedersen SB, Moller-Pedersen T, Bak JF . Regional differences in triglyceride breakdown in human adipose tissue: effects of catecholamines, insulin, and prostaglandin E2 Metabolism 1991 40: 990–996.

    CAS  Article  Google Scholar 

  39. 39

    Mauriege P, Galitzky J, Berlan M, Lafontan M . Heterogeneous distribution of beta and alpha-2 adrenoceptor binding sites in human fat cells from various fat deposits: functional consequences Eur J Clin Invest 1987 17: 156–165.

    CAS  Article  Google Scholar 

  40. 40

    Xu X, De Pergola G, Björntorp P . Steroid hormone effects on adipose tissue growth and metabolism. In: Crepaldi G, Tiengo A, Enzi G (eds) Diabetes, obesity and hyperlipidemias IV: Proceedings of the 5th European Symposium on Metabolism. Elsevier: Amsterdam, 1990, pp 173–178.

    Google Scholar 

  41. 41

    Rebuffé-Scrive M, Mårin P, Björntorp P . Effect of testosterone on abdominal adipose tissue in men Int J Obes 1991 15: 791–795.

    PubMed  Google Scholar 

  42. 42

    Mårin P, Odén B, Björntorp P . Assimilation and mobilization of triglycerides in subcutaneous abdominal and femoral adipose tissue in vivo in men: effects of androgens J Clin Endocrinol Metab 1995 80: 239–243.

    PubMed  Google Scholar 

  43. 43

    Sjogren J, Li M, Björntorp P . Androgen hormone binding to adipose tissue in rats Biochim Biophys Acta 1995 1244: 117–120.

    CAS  Article  Google Scholar 

  44. 44

    Boyko EJ, Leonetti DL, Bergstrom RW, Newell-Morris L, Fujimoto WY . Visceral adiposity, fasting plasma insulin, and blood pressure in Japanese-Americans Diabetes Care 1995 18: 174–181.

    CAS  Article  Google Scholar 

  45. 45

    Boyko EJ, Leonetti DL, Bergstrom RW, Newell-Morris L, Fujimoto WY . Visceral adiposity, fasting plasma insulin, and lipid and lipoprotein levels in Japanese Americans Int J Obes 1996 20: 801–808.

    CAS  Google Scholar 

  46. 46

    Randle PJ, Garland PB, Hales CN, Newsholme EA . The glucose-fatty acid cycle: its role in insulin sensitivity and the metabolic disturbances of diabetes mellitus Lancet 1963 i: 785–789.

    Article  Google Scholar 

  47. 47

    Boden G . Role of fatty acids in he pathogenesis of insulin resistance and NIDDM Diabetes 1997 46: 3–10.

    CAS  Article  Google Scholar 

  48. 48

    Boyko EJ, Leonetti DL, Bergstrom RW, Newell-Morris L, Fujimoto WY . Low insulin secretion and high fasting insulin and C-peptide levels predict increased visceral adiposity Diabetes 1996 45: 1010–1015.

    CAS  Article  Google Scholar 

  49. 49

    Tchernof A, Després JP, Dupont A, Belanger A, Nadeau A, Prud'homme D, Moorjani S, Lupien PJ, Labrie F . Relation of steroid hormones to glucose tolerance and plasma insulin levels in men Diabetes Care 1995 18: 292–299.

    CAS  Article  Google Scholar 

  50. 50

    Haffner SM, Valdez RA, Morales PA, Hazuda HP, Stern MP . Decreased sex hormone-binding globulin predicts non-insulin dependent diabetes mellitus in women but not in men J Clin Endocrinol Metab 1993 77: 56–60.

    CAS  PubMed  Google Scholar 

  51. 51

    Laakso M . How good a marker is insulin level for insulin resistance? Am J Epidemiol 1993 137: 959–965.

    CAS  Article  Google Scholar 

Download references


This work was supported by NIH grants DK 31170, HL 49293, RR 00037, DK 35816 and DK 17047. Dr Tsai is supported by a Health Services fellowship from the Health Services Research and Development, US Department of Veterans Affairs. We wish to thank Ms Jane Shofer for her skilled technical support in maintaining the JACDS database. We are grateful to the King County Japanese American Community for their support and cooperation

Author information



Corresponding author

Correspondence to EC Tsai.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Tsai, E., Boyko, E., Leonetti, D. et al. Low serum testosterone level as a predictor of increased visceral fat in Japanese-American men. Int J Obes 24, 485–491 (2000). https://doi.org/10.1038/sj.ijo.0801183

Download citation


  • testosterone
  • visceral adiposity
  • body fat distribution
  • diabetes mellitus
  • insulin resistance
  • Japanese Americans
  • computed tomography

Further reading


Quick links