Differences in Insulin Suppression of Free Fatty Acid Levels by Gender and Glucose Tolerance Status
Relation to Plasma Triglyceride and Apolipoprotein B Concentrations
Most discussions of relations of insulin resistance to coronary heart disease risk factors have focused on insulin-stimulated glucose uptake, but insulin suppression of plasma free fatty acid (FFA) levels is also important in lipid and lipoprotein metabolism. To identify groups with impaired insulin suppression of FFAs, we studied a multiethnic cohort of 1521 women and men at four US centers that comprise the Insulin Resistance Atherosclerosis Study (IRAS): 682 with normal glucose tolerance, 352 with impaired glucose tolerance, and 487 with non–insulin-dependent diabetes. The FFA level 2 hours after a 75-gm oral glucose load adjusted for fasting FFAs was used as the measure of insulin suppression. After adjustment for age, center, ethnicity, body mass index, and fasting and 2-hour insulin levels, 2-hour FFA levels were significantly higher in men than women and in persons with impaired glucose tolerance and non–insulin-dependent diabetes mellitus versus normal glucose tolerance. The gender difference was largely accounted for by differences in central obesity as measured by waist-hip ratio; the difference by glucose tolerance status was not affected by central obesity, suggesting a different mechanism. In multivariate regression analyses, 2-hour FFA levels were strongly related to fasting triglyceride and apoB levels, respectively, after adjustment for age, fasting and 2-hour insulin concentrations, and fasting FFA concentrations. In summary, elevated plasma apoB and triglyceride concentrations associated with male gender and with glucose intolerance are partly accounted for by differences in the ability of insulin to suppress FFA concentrations.
- Received August 9, 1995.
- Revision received May 13, 1996.
Although discussions of the role of insulin resistance in CHD risk have generally focused on the action of insulin in stimulating glucose uptake,1 2 3 insulin has another important function that affects CHD risk factors, namely, the regulation of lipolysis.4 5 6 7 Insulin is the major hormone to inhibit hydrolysis of TGs in adipose cells into glycerol and FFAs.8 Together with glucose, insulin may also play a role in the reesterification of FFAs in adipose cells, promoting TG storage.9 By these mechanisms, insulin lowers plasma FFA levels.
Circulating plasma FFAs are the major substrates for TG synthesis in the liver,5 and they stimulate apoB secretion from the liver.10 11 12 Thus, the ability of insulin to suppress plasma FFA concentrations plays a major role in hepatic VLDL TG synthesis and secretion, apoB secretion,10 11 12 and plasma VLDL cholesterol and apoB concentrations.5 6 7 13
A defect in insulin suppression of FFAs has been shown in patients with NIDDM4 14 15 16 and has been proposed to account in part for the dyslipidemia associated with NIDDM.6 Since persons with IGT have marked resistance to insulin-stimulated glucose uptake,17 18 they may also have a defect in insulin suppression of FFAs. However, the results of two previous studies examining this issue have been contradictory.19 20
Compared with women, men have a deficit in insulin suppression of plasma FFA levels after an oral glucose load.20 21 We have shown21 in NGT individuals from two ethnic groups in the United Kingdom that lower TG and apoB concentrations in women could be explained in part by the fact that after an oral glucose challenge women suppress plasma FFA concentrations more than men. Since some of the sex difference in CHD risk can be attributed to sex differences in lipid and lipoprotein levels,22 sex differences in FFA suppression may in part explain sex differences in CHD risk.
This study is the first large-scale, multiethnic epidemiological study to examine the effects of both gender and glucose tolerance status on insulin suppression of FFA concentrations. Data are from a cohort of persons with NGT, IGT, and NIDDM in four centers in the United States that comprise the ongoing Insulin Resistance Atherosclerosis Study (IRAS).
Details of the study methodology for the IRAS are available.23 Briefly, 1626 non-Hispanic white (38%), African American (29%), and Hispanic (34%) men and women aged 45 to 69 years were enrolled at four US clinical centers from 1992 through 1994. At the Los Angeles and Oakland, Calif, sites, subjects were recruited from a large health maintenance organization membership. At San Luis Valley, Colo, and San Antonio, Tex, subjects were recruited from cohorts in ongoing population-based epidemiolgical surveys. There were 720 subjects with NGT, 369 with IGT, and 537 with NIDDM according to World Health Organization criteria24 in the four centers. Subjects who had ever taken insulin were ineligible for the study. One hundred five subjects had missing data for either the fasting or 2-hour FFA concentration and were excluded from the analyses. This left 1521 subjects: 682 with NGT, 352 with IGT, and 487 with NIDDM.
Protocol and Laboratory Methods
The examinations were conducted at two clinic visits ≈1 week apart. All subjects fasted for 12 hours and refrained from heavy exercise, smoking, and alcohol consumption for 24 hours before each visit, which began at approximately 8 am and lasted 4 hours. Subjects with NIDDM on oral hypoglycemic agents held their am dose the morning of the studies.
Subjects were weighed to the nearest 0.1 kg in light clothing without shoes. Height was recorded to the nearest 0.5 cm, and BMI was calculated as weight in kilograms divided by height in meters squared. Minimum waist girth was measured in duplicate to the nearest 0.5 cm with a steel tape at the natural indentation between the 10th rib and the iliac crest at end expiration. The hip was measured at the greatest protrusion of the buttocks. The average of duplicate measurements was used to calculate the WHR.
On the first morning, blood was drawn from seated subjects for measurement of fasting glucose, insulin, FFA, and lipid and lipoprotein concentrations. A 75-g oral glucose load was administered, and blood was drawn 2 hours later for repeat measurements of glucose, insulin, and FFAs. At the second visit, resistance to insulin-stimulated glucose uptake was assessed by using the frequently sampled intravenous glucose tolerance test.25 26 Details of this protocol as used in IRAS are available.18 An Si (expressed as minutes per microunit per milliliter) was calculated by using a modified version of the program MINMOD.25 Higher Si values indicate greater sensitivity to insulin-stimulated glucose uptake.
Plasma glucose concentrations were measured in duplicate by using the glucose oxidase method.27 The intra-assay CV was 1.9% for low and 0.7% for high concentration pools; the interassay CV was 3.2% for low and 2.3% for high concentration pools. Plasma insulin concentrations (all control interassay CVs <11%) were determined in duplicate by using a dextran-charcoal radioimmunoassay28 according to 1986 World Health Organization international standards. Insulin intra-assay CVs were 10.6% for low and 10.2% for high concentration pools; interassay CVs were 19.3% for low and 17.4% for high concentration pools. Plasma FFA concentrations were measured colorimetrically (interassay CV, 8%).29 Fasting plasma TG concentrations, expressed as the average of values determined on days 1 and 2, were determined by using enzymatic methods30 (interassay CV, 3.6%). Plasma total apoB concentrations were determined by using an immunoprecipitation technique (SPQ kit from Instar, Inc) (interassay CV, 4.1%).
Data were analyzed by using the SAS GLM procedure. Regression models were constructed to test the following three hypotheses: 2-hour FFA concentrations are lower in women than men, with relative suppression of FFAs greater in women; persons with IGT and NIDDM have higher fasting and 2-hour FFA concentrations than those with NGT; and the 2-hour FFA concentration is strongly associated with plasma TG and apoB concentrations in all groups. Age and indices of body fatness were entered into the models since they are known to affect the dependent variables of interest. Because of the four-site, multiethnic study design, all models were controlled for effects of center and ethnicity.
FFA suppression was assessed by using models of 2-hour FFAs that adjusted for fasting FFAs by including fasting FFA level as an independent variable. This method allows a precise estimation in the model of the relative change between fasting and 2-hour FFA levels. This is more powerful than analyzing a predetermined relationship in the form of a function of these two variables, eg, percent suppression or fasting FFA minus 2-hour FFA levels.
Because of skewed distributions, natural log transformations were used for TG, fasting and 2-hour insulin, and FFA concentrations to meet assumptions of regression. Because the Si included zero values, the natural log of (Si+1) was used in models including Si as covariate in normal subjects. The highly skewed distribution of Si and the high proportion of zero values among subjects with NIDDM necessitated using quintiles of Si for analyses with all glucose tolerance groups combined. Subjects with zero values for Si were included in all the models.
Results are shown in the tables as β coefficients, SEs of the β coefficients, and probability values. A direct interpretation of the β coefficient is as follows. For an independent variable, eg, age, for every 1-year increase there is a corresponding increase in the dependent variable, eg, the natural log of fasting FFA concentrations. The magnitude of this increase is equal to the β coefficient given constant levels of the other independent variables in the model, eg, insulin and gender. Because of the need to use log transformations of variables in the models to meet the distributional assumptions of regression methods, some ease of interpretability is sacrificed. Therefore, the information obtained from these β coefficients includes (1) the direction of the relationships between the dependent and independent variables, (2) the magnitude of the β coefficient relative to its standard error, (3) the level of significance of the coefficients, and (4) the magnitude of the β coefficient in one model relative to another.
Anthropometric and metabolic data for the subjects are shown in Table 1⇓. Age, BMI, WHR, and fasting insulin, fasting and 2-hour glucose, FFA, TG, and apoB concentrations all increased with glucose tolerance status from NGT to NIDDM. Two-hour insulin concentrations increased from NGT to IGT but were lower in those with NIDDM than IGT. Si decreased as expected from NGT to IGT and NIDDM.
Within each glucose tolerance group, women were similar to men in age but had a higher BMI and lower WHR. Women had slightly lower fasting glucose but similar 2-hour glucose concentrations compared with men. Women and men had equivalent fasting insulin levels in each group, but 2-hour insulin was higher in women versus men in the NGT and NIDDM groups and lower in women versus men in the IGT group. Within each glucose tolerance group, women had higher fasting FFA, lower 2-hour FFA, and lower TG concentrations than men. Women in the NGT and IGT groups had lower apoB levels than men, but apoB concentrations were higher in women in the NIDDM group.
Determinants of Fasting and 2-Hour FFAs in NGT Subjects
Regression analyses of fasting and 2-hour plasma FFA concentrations on anthropometric and metabolic variables in persons with NGT are shown in Table 2⇓ in a series of five models. Women had higher fasting FFA concentrations than men after adjustment for age, fasting insulin, or Si (models 1 and 2). Neither BMI nor WHR contributed significantly to fasting FFA concentrations (P=.23 and 0.34, respectively; not shown).
Model 1 shows that fasting plasma FFA concentrations increased with fasting insulin concentrations. When Si was entered into the model (model 2), the effect of fasting insulin was reduced to insignificance, indicating that the relation of insulin concentration to FFAs is due to the correlation of insulin level with insulin sensitivity. All subsequent models were constructed by using either fasting insulin levels or Si but not both variables together. Si and fasting insulin levels contributed very similar information in the models; use of one versus the other had no effect on the comparisons of interest, namely, gender and glucose tolerance differences in FFA concentrations. Therefore, final models are shown with fasting insulin levels rather than Si.
Model 3 shows that women with NGT had significantly lower 2-hour FFA concentrations than men after adjustment for age, fasting FFA concentrations, and fasting and 2-hour insulin concentrations. Adding BMI to the model did not affect the gender difference (not shown).
Model 4 shows that the relationship of fasting insulin concentrations to 2-hour FFA levels is partly explained by the Si, again suggesting that the relationship of fasting insulin to 2-hour FFAs is due to the correlation of fasting insulin level to insulin sensitivity. Models 3 and 4 also show that at a given fasting insulin or Si level, the higher the 2-hour insulin, the lower the 2-hour FFAs. Thus, individuals with a given degree of insulin resistance (as reflected in the fasting insulin level or Si) will vary in their insulin secretory response to a glucose challenge (as reflected in the 2-hour insulin level). Therefore, at a given degree of insulin resistance, the higher the 2-hour insulin level, the more the 2-hour FFA level will be suppressed.
Addition of Si to the model did not diminish the gender difference in FFA suppression, indicating that this particular gender difference is not due to gender differences in insulin sensitivity.
While BMI did not alter the gender difference in 2-hour FFA concentrations, WHR reduced it to insignificance (compare models 3 and 5).
Determinants of Fasting and 2-Hour FFAs in Subjects With NGT, IGT, and NIDDM
Regressions of FFA concentrations on anthropometric and metabolic variables for all three glucose tolerance groups combined are shown in Table 3⇓ in a series of four models. NGT women in the combined group had higher fasting FFA concentrations than men. Age, fasting insulin, gender, and diabetes status all contributed to fasting FFA concentrations (model 1).
Although BMI was not significantly associated with fasting FFAs in persons with NGT, it was highly significantly associated with fasting FFAs when all subjects were combined (Table 3⇑, model 2). However, the interaction term for BMI with glucose tolerance status was not significant (not shown), indicating that the relation of BMI to fasting FFAs was not significantly different for any of the three groups. When WHR was substituted for BMI in the model, it was not significantly associated with fasting FFAs (P=.59), nor did it affect the gender difference in fasting FFAs (not shown).
After adjustment for age, center, ethnicity, BMI, and fasting insulin level, fasting FFA concentrations were higher in both women and men with IGT and NIDDM compared with those with NGT (Figure⇓). There were no significant interactions of glucose tolerance status with gender in any model (not shown). This indicates that the effect of glucose tolerance status on fasting FFA concentrations did not differ between women and men, ie, fasting FFAs increased similarly from NGT to IGT to NIDDM in both women and men.
Regressions of 2-hour FFA concentrations on anthropometric and metabolic variables for all three glucose tolerance groups are shown in models 3 and 4 of Table 3⇑. Men had higher 2-hour FFA concentrations than women after adjustment for age, fasting FFAs, fasting and 2-hour insulin, and glucose tolerance status (model 3). BMI was significantly associated with 2-hour FFA concentrations (P=.002; not shown) but did not affect the gender difference. However, when WHR was substituted for BMI in the model, the gender difference in 2-hour FFA concentrations was substantially reduced (compare models 3 and 4).
Models 3 and 4 also show that fasting insulin was directly associated with 2-hour FFA concentrations, while 2-hour insulin was inversely associated; both relations were highly significant. One interpretation is that insulin-resistant subjects with higher fasting insulin levels are resistant to suppression of FFAs as well as glucose uptake. For a given degree of insulin resistance, as measured by fasting insulin concentration, the higher the 2-hour insulin, the greater the degree of FFA suppression.
Both women and men with IGT or NIDDM had higher 2-hour FFA concentrations than those with NGT after adjustment for fasting FFAs, age, center, ethnicity, BMI, and 2-hour insulin level. There was no significant interaction of glucose tolerance status with gender in any model, indicating no difference between men and women in the effect of glucose tolerance status on 2-hour FFA concentrations.
TGs and ApoB
Regression analyses of fasting plasma TG concentrations on anthropometric and metabolic variables are shown in Table 4⇓ in a series of six models. Models 1 through 3 are for persons with NGT. After adjustment for age, there was a highly significant gender difference in fasting TG concentrations (model 1). After adding fasting and 2-hour FFA concentrations to the model (model 2), the gender effect was reduced by about one quarter, though it was still highly significant. Of note, 2-hour FFA concentrations contributed far more than fasting FFA levels to TG concentrations, suggesting that it is the ability of insulin to suppress FFAs, rather than the absolute FFA level, that affects TG concentration. Model 3 shows that fasting and 2-hour insulin concentrations made highly significant contributions to TG levels but did not alter the gender difference.
Models 4 through 6 (Table 4⇑) combine all three glucose tolerance groups. Relationships were similar to those seen for NGT subjects.
Regressions of plasma apoB concentrations on other metabolic variables are shown in Table 5⇓ in a series of six models. Women with NGT had lower apoB concentrations than men after adjustment for age (model 1). The gender difference in apoB concentrations was reduced by adding fasting and 2-hour FFA concentrations to the model (model 2); only 2-hour FFAs contributed significantly. Model 3 shows that while fasting and, to a lesser extent, 2-hour insulin concentrations contributed significantly to apoB concentrations, they had no effect on the gender difference.
When the three glucose tolerance groups were combined (Table 5⇑, models 4 through 6), gender was no longer significantly associated with apoB concentrations because among those with NIDDM, women had higher apoB concentrations than men, whereas in the other two groups, women had lower apoB concentrations (Table 1⇑). As in the NGT group, fasting and 2-hour insulin and 2-hour FFA levels were highly significantly associated with apoB concentrations when all three glucose tolerance groups were combined.
There are three main findings in the current study. First, the data demonstrate that women in all three glucose tolerance groups have greater FFA suppression than men 2 hours after an oral glucose load. This replicates earlier studies of women versus men with NGT20 21 31 32 and extends the findings to women versus men with IGT and NIDDM. The gender difference was independent of age, BMI, fasting and 2-hour insulin concentrations, and glucose tolerance status, but was largely explained by the gender difference in WHR. In vitro, intra-abdominal fat cells are more resistant to insulin-mediated suppression of lipolysis than are subcutaneous fat cells.33 34 In vivo, central obesity is associated with impaired suppression of FFAs by insulin in women35 and men.36 The greater proportion of body fat stored intra-abdominally in men37 may explain their relative impairment in FFA suppression.
Second, the data demonstrate a defect in insulin suppression of FFAs in persons with IGT and NIDDM compared with those with NGT. “Insulin resistance” in persons with IGT, therefore, encompasses both resistance to insulin-stimulated glucose uptake17 18 and FFA suppression. This finding is in contrast to the model previously proposed,6 38 ie, that in the continuum from NGT to IGT and NIDDM, FFA levels begin to rise only after the onset of β cell failure and relative hypoinsulinemia.
Insulin clamp studies investigating mechanisms for impaired FFA suppression among persons with NIDDM have shown defects in several areas of FFA metabolism: FFA turnover, nonoxidative FFA disposal (an estimate of FFA reesterification), and FFA oxidation.16 In one study, absolute nonoxidative and oxidative FFA metabolism were actually higher in NIDDM patients than control subjects and were strongly correlated with plasma FFA concentrations.16 This suggests that the major defect in FFA metabolism in NIDDM is impaired insulin suppression of lipolysis (production) rather than impaired disposal.
While the impairments in FFA suppression associated in this study with IGT and NIDDM were accompanied by impairments in insulin-stimulated glucose uptake, the gender difference in FFA suppression was not (Table 1⇑). Furthermore, while adding WHR to the models greatly diminished the gender difference in FFA suppression, it did not affect differences by glucose tolerance status. These findings, and the fact that both glucose tolerance status and gender independently contributed to FFA suppression (Table 3⇑, model 4), suggest different mechanisms for the effects of gender and glucose tolerance status on insulin suppression of plasma FFA levels.
The third finding of this study is that 2-hour FFA concentrations are highly significantly associated with total apoB concentrations. The relations of FFA to apoB concentrations that we found in the current study are consistent with the stimulation of apoB secretion from the liver by FFAs as has been demonstrated by using cultured HepG2 cells.39
Among those with NGT, 2-hour FFA concentrations explained a statistically significant portion of the gender differences in TG and apoB levels. This last finding is notable given that 2-hour FFA concentration is only a crude measure of insulin suppression of lipolysis. FFA concentrations 2 hours after an oral glucose load40 or mixed meal41 represent the point of maximal suppression. But because the ability of insulin to suppress FFA concentrations is reflected in FFA levels at numerous times throughout the day, day-long postprandial FFA levels are likely to be far more important determinants of plasma TG and apoB concentrations than are maximally suppressed FFA levels. Had we obtained FFA levels at additional times or measured the rate of fall in FFAs, we may have been able to explain more of the variation in TG and apoB concentrations. One clamp study showed that the slope of decay in FFA levels in response to insulin infusion explained 66% of the variation in fasting plasma TG concentrations in men with hypertriglyceridemia and control subjects.7
Whereas others have proposed that a sex difference in sensitivity to insulin-stimulated glucose uptake may be partly responsible for the sex difference in CHD risk,42 we propose instead that a sex difference in insulin suppression of FFAs may play the more important role. In a previous study relating sex differences in FFA suppression to TG and apoB concentrations among subjects with NGT, we speculated21 that women with NIDDM may lose the advantageous lipoprotein profile associated with female gender because of an impairment in insulin suppression of FFAs, ie, that women and men with NIDDM would have equally impaired FFA suppression. We did not demonstrate this in the current study. Although NIDDM women had impaired FFA suppression compared with IGT and NGT women, men with NIDDM had an even greater impairment. However, it is still possible that women with NIDDM have defects in FFA suppression that are not apparent at the maximally suppressed level. Had we measured day-long FFA levels or FFAs at earlier times after the glucose challenge, a relatively greater defect in FFA suppression in diabetic women may have been apparent. Women compared with men with NIDDM are reported to have impairment of FFA suppression at 30 but not 120 minutes after an oral glucose challenge.20
One further interesting result of this study is the significantly higher fasting FFA concentration in women than men. This may be due to the fact that during periods of low insulin levels, such as occur in the fasting state, other factors, such as gender differences in percent total body fat or effects of lipolytic hormones (eg, cortisol or catecholamines) are more important determinants of FFA concentration.
In summary, we found that men compared with women and persons of both genders with IGT and NIDDM compared with those with NGT have impaired insulin-mediated FFA suppression after an oral glucose load. The defects in FFA suppression due to gender and glucose tolerance status appear to be mediated by different mechanisms. Defects in insulin suppression of FFA concentrations are strongly related to lipid and lipoprotein risk factors for CHD.
Selected Abbreviations and Acronyms
|BMI||=||body mass index|
|CHD||=||coronary heart disease|
|CV||=||coefficient of variation|
|FFA||=||free fatty acid|
|IGT||=||impaired glucose tolerance|
|NGT||=||normal glucose tolerance|
|NIDDM||=||non–insulin-dependent diabetes mellitus|
|Si||=||insulin sensitivity index|
This work was supported in part by National Institutes of Health grants HL-47902, HL-47887, HL-08506, DK-29867, and T32-AG-00093.
Laakso M, Sarlund H, Salonen R, Suhonen M, Pyorala K, Salonen JT, Karhapaa P. Asymptomatic atherosclerosis and insulin resistance. Arterioscler Thromb. 1991;11:1068-1076.
Bierman EL, Dole VP, Roberts TN. An abnormality of nonesterified fatty acid metabolism in diabetes mellitus. Diabetes. 1957;6:475-479.
Reaven GM, Greenfield MS. Diabetic hypertriglyceridemia: evidence for three clinical syndromes. Diabetes. 1981;30(suppl 2):66-75.
Yki-Jarvinen H, Taskinen M-R. Interrelationships among insulin's antilipolytic and glucoregulatory effects and plasma triglycerides in nondiabetic and diabetic patients with endogenous hypertriglyceridemia. Diabetes. 1988;37:1271-1278.
Wolfe RR, Peters EJ. Lipolytic response to glucose infusion in human subjects. Am J Physiol. 1987;252:E218-E223.
Pullinger CR, North JD, Teng BB, Rifici VA, Ronhild de Brito AE, Scott J. The apolipoprotein B gene is constitutively expressed in HepG2 cells: regulation of secretion by oleic acid, albumin, and insulin, and measurement of the mRNA half-life. J Lipid Res. 1989;30;1065-1077.
Dixon JL, Furukawa S, Ginsberg HN. Oleate stimulates secretion of apolipoprotein B-containing lipoproteins from Hep G2 cells by inhibiting early intracellular degradation of apolipoprotein B. J Biol Chem. 1991;266:5080-5086.
Cianflone K, Dahan S, Monge JC, Sniderman AD. Pathogenesis of carbohydrate-induced hypertriglyceridemia using HepG2 cells as a model system. Arterioscler Thromb. 1992;12:271-277.
Kissebah AH, Alfarsi S, Evans DJ, Adams PW. Integrated regulation of very-low density triglyceride and apolipoprotein-B kinetics in non-insulin-dependent diabetes mellitus. Diabetes. 1982;31:217-225.
Groop LC, Bonadonna RC, DelPrato S, Ratheiser K, Zyck K, Ferrannini E, DeFronzo RA. Glucose and free fatty acid metabolism in non-insulin-dependent diabetes mellitus. J Clin Invest. 1989;84:205-213.
Saad MF, Anderson RL, Laws A, Watanabe RM, Kades WW, Chen Y-DI, Sands RE, Pei D, Savage PJ, Bergman RN, for the Insulin Resistance Atherosclerosis Study. A comparison between the minimal model and the glucose clamp in the assessment of insulin sensitivity across the spectrum of glucose tolerance. Diabetes. 1994;431:1114-1121.
Byrne CD, Wareham NJ, Brown DC, Clark PMS, Cox LJ, Day NE, Palmer CR, Wang TWM, Williams DRR, Hales CN. Hypertriglyceridemia in subjects with normal and abnormal glucose tolerance: relative contributions of insulin secretion, insulin resistance and suppression of plasma non-esterified fatty acids. Diabetologia. 1994;37:889-896.
McKeigue PM, Laws A, Chen Y-DI, Marmot MG, Reaven GM. Relation of plasma triglyceride and apolipoprotein B levels to insulin-mediated suppression of nonesterified fatty acids: possible explanation for sex differences in lipoprotein pattern. Arterioscler Thromb. 1993;8:1187-1192.
World Health Organization. Diabetes Mellitus: Report of a WHO Study Group. Geneva, Switzerland; 1985. Technical report series No. 727.
Bergman RN, Prager R, Volund A, Olefsky JM. Equivalence of the insulin sensitivity index in man derived by the minimal model and the euglycemic glucose clamp. J Clin Invest. 1987;79:790-800.
Kadish AH, Litle RL, Sternberg JC. A new and rapid method for determination of glucose by measurement of rate of oxygen consumption. Clin Chem. 1968;14:116-131.
Noma A, Okage H, Kita M. A new colorimetric micro-determination of free fatty acids in serum. Clin Chim Acta. 1972;43:317-320.
Wahlfeld AW. Triglycerides: determination after enzymatic hydrolysis. In: Bergmeyer HV, ed. Methods of Enzymatic Analysis. New York, NY: Academic Press; 1974:1831-1835.
Boyns DR, Crossley JN, Abrams ME, Jarrett RJ, Keen H. Oral glucose tolerance and related factors in a normal population sample, I: blood sugar, plasma insulin, glyceride, and cholesterol measurements and the effects of age and sex. BMJ. 1969;1:595-598.
Bolinder J, Kaer L, Ostman J, Arner P. Differences at the receptor and post-receptor levels between human omental and subcutaneous adipose tissue in the action of insulin on lipolysis. Diabetes. 1983;32:117-123.
Jensen MD, Haymond MW, Rizza RA, Cryer PE, Miles JM. Influence of body fat distribution on free fatty acid metabolism in obesity. J Clin Invest. 1989;83:1168-1173.
Seidell JC, Oosterlee A, Thijssen MAO, Burema J, Deurenberg P, Hautvast JGAJ, Ruijs JHJ. Assessment of intra-abdominal and subcutaneous abdominal fat: relation between anthropometry and computed tomography. Am J Clin Nutr. 1987;45:7-13.
Dixon JL, Ginsberg HN. Regulation of hepatic secretion of apolipoprotein B-containing lipoproteins: information obtained from cultured liver cells. J Lipid Res. 1993;34:167-179.
Laws A, Jeppesen JL, Maheux PC, Schaaf P, Chen Y-DI, Reaven GM. Resistance to insulin-stimulated glucose uptake and dyslipidemia in Asian Indians. Arterioscler Thromb. 1994;14:917-922.
Donahue RP, Orchard TJ, Becker DJ, Kuller LH, Drash AL. Sex differences in the coronary heart disease risk profile: a possible role for insulin. Am J Epidemiol. 1987;125:650-657.