A genome-wide scan for human obesity genes reveals a major susceptibility locus on chromosome 10.

Obesity, a common multifactorial disorder, is a major risk factor for type 2 diabetes, hypertension and coronary heart disease (CHD). According to the definition of the World Health Organization (WHO), approximately 6-10% of the population in Westernized countries are considered obese. Epidemiological studies have shown that 30-70% of the variation in body weight may be attributable to genetic factors. To date, two genome-wide scans using different obesity-related quantitative traits have provided candidate regions for obesity. We have undertaken a genome-wide scan in affected sibpairs to identify chromosomal regions linked to obesity in a collection of French families. Model-free multipoint linkage analyses revealed evidence for linkage to a region on chromosome 10p (MLS=4.85). Two further loci on chromosomes 5cen-q and 2p showed suggestive evidence for linkage of serum leptin levels in a genome-wide context. The peak on chromosome 2 coincided with the region containing the gene (POMC) encoding pro-opiomelanocortin, a locus previously linked to leptin levels and fat mass in a Mexican-American population and shown to be mutated in obese humans. Our results suggest that there is a major gene on chromosome 10p implicated in the development of human obesity, and the existence of two further loci influencing leptin levels.

Melanocortin-4 receptor mutations are a frequent and heterogeneous cause of morbid obesity.

By integrating an agonist satiety signal, provided by alpha-melanocyte-stimulating hormone (alpha-MSH), and an antagonist signal, provided by agouti-related protein (AGRP), the melanocortin-4 receptor (MC4-R) is a key element in the hypothalamic control of food intake. Inactivation of the gene encoding this G protein-coupled receptor causes obesity in mice. In humans, frameshift mutations in MC4-R cause an early-onset dominant form of obesity in two families. In this study we find a high frequency (4%) of rare heterozygous MC4-R mutations in a large population of morbidly obese patients. No such mutations were found in controls. By analyzing the phenotypes of the probands carrying these mutations, we demonstrate that these patients display a common, nonsyndromic form of obesity. Interestingly, functional analysis of the mutant receptors indicates that obesity-associated defects in MC4-R range from loss of function to constitutive activation. Transmission of these mutations in the families of the carriers indicates a variable expressivity that is not related to the functional severity of the mutations. This variable expressivity of MC4-R-associated obesity is not due to variations in genes for alpha-MSH or AGRP. Taken together, these results demonstrate that MC4-R mutations are a frequent but heterogeneous genetic cause of morbid obesity.

Soluble leptin receptor in serum of subjects with complete resistance to leptin: relation to fat mass.

Leptin resistance and obesity have been related to mutations of the leptin receptor gene in rodents and, recently, in a consanguineous family. The latter mutation results in a receptor lacking transmembrane and intracellular domains. Homozygous and heterozygous individuals with this mutation had serum leptin levels higher than expected, given their BMIs: 600, 670, and 526 ng/ml and 145, 362, 294, 240, and 212 ng/ml, respectively. Their serum leptin was fractionated by gel filtration: >80% was present as a high-molecular size complex vs. 7.5% in the nonmutated sister. Western blot analysis showed a band at 146 kDa reacting specifically with an antibody directed against the leptin receptor ectodomain. In 10 obese control subjects, as in the mutated patients, free leptin levels correlated with BMI (r = 0.70, P = 0.0011) and reflected fat mass, regardless of leptin receptor functioning. In the patients, bound leptin levels correlated with BMI (r = 0.99, P = 0.0002) and were related to the number of mutated alleles. These data demonstrate that the truncated receptor is secreted into blood and binds the majority of serum leptin, markedly increasing bound and total leptin. Free serum leptin was similarly correlated with BMI in the mutated and nonmutated obese individuals, providing evidence that the relationship between BMI and circulating free leptin is preserved in this family. This finding suggests that the leptin receptor itself may not be specifically involved in the control of leptin secretion, and it supports the concept of relative resistance to leptin in common obesity.

A sib-pair analysis study of 15 candidate genes in French families with morbid obesity: indication for linkage with islet 1 locus on chromosome 5q.

As part of an ongoing search for susceptibility genes in obese families, we performed linkage analyses in 101 French families between qualitative and quantitative traits related to morbid obesity and polymorphisms located in or near 15 candidate genes whose products are involved in body weight regulation. These included cholecystokinin A and B receptors (CCK-AR and CCK-BR), glucagon-like peptide 1 receptor (GLP-1R), the LIM/homeodomain islet-1 gene (Isl-1), the caudal-type homeodomain 3 (CDX-3), the uncoupling protein 1 (UCP-1), the beta3-adrenoceptor (beta3-AR), the fatty acid-binding protein 2 (FABP-2), the hormone-sensitive lipase (HSL), the lipoprotein lipase (LPL), the apoprotein-C2 (apo-C2), the insulin receptor substrate-1 (IRS-1), the peroxisome proliferator-activated receptor-gamma (PPAR-gamma), tumor necrosis factor-alpha (TNF-alpha), and the liver carnitine palmitoyltransferase-1 (CPT-1). Phenotypes related to obesity such as BMI, adult life body weight gain, fasting leptin, insulin, fasting glycerol, and free fatty acids were used for nonparametric sib-pair analyses. A weak indication for linkage was obtained between the Isl-1 locus and obesity status defined by a z score over one SD of BMI (n = 226 sib pairs, pi = 0.54 +/- 0.02, P = 0.03). Moreover, a suggestive indication for linkage was found between the Isl-1 locus and BMI and leptin values (P = 0.001 and 0.0003, respectively) and leptin adjusted for BMI (P = 0.0001). Multipoint analyses for leptin trait with Isl-1 and two flanking markers (D5S418 and D5S407) showed that the logarithm of odds (LOD) score is 1.73, coinciding with the Isl-1 locus. Although marginally positive indications for linkage in subgroups of families were found with IRS-1, CPT-1, and HSL loci, our data suggested that these genes are not major contributors to obesity. Whether an obesity susceptibility gene (Isl-1 itself or another nearby gene) lies on chromosome 5q should be determined by further analyses.

Indication for linkage of the human OB gene region with extreme obesity.

Obesity is one of the most significant risk factors for hypertension, coronary heart disease, and NIDDM (Frayn KN, Coppack SW: Insulin resistance, adipose tissue and coronary heart disease. Clin Sci 82:1-8, 1992; Kaplan NM: The deadly quartet: upper-body obesity, glucose intolerance, hypertriglyceridemia, and hypertension. Arch Intern Med 149:1514-1520, 1989). While family segregation, adoption, and twin studies have indicated that degree of adiposity has a significant genetic component (Stunkard AJ, Harris JR, Pedersen NL, McClearn GE: The body-mass index of twins who have been reared apart. N Engl J Med 322:1483-1487, 1990; Bouchard C, Despres J-P, Mauriege P: Genetic and nongenetic determinants of regional fat distribution. Endocr Rev 14:72-93, 1993), the genes and predisposing mutations remain poorly understood. This is in contrast to several well-defined genetic models for obesity in rodents, particularly the mouse obese (ob) gene, in which loss-of-function mutations cause severe obesity. Recent studies have demonstrated a substantial reduction in body fat when recombinant ob protein (leptin) is administered to mice. To test the relevance of these observations to human obesity, the location of the human homologue (OB) was established by radiation hybrid mapping and eight microsatellite markers spanning the OB gene region (7q3l.3) were genotyped in 101 obese French families. Affected-sib-pair analyses for extreme obesity, defined by BMI >35 kg/m2, revealed suggestive evidence for linkage to three markers located within 2 cM of the OB gene (D7S514, D7S680, and D7S530). The OB gene is therefore a candidate for genetic predisposition to extreme obesity in a subset of these families.

Genetic studies of the sulfonylurea receptor gene locus in NIDDM and in morbid obesity among French Caucasians.

The sulfonylurea receptor (SUR) is a key component in glucose-stimulated insulin secretion. Obesity and NIDDM are frequently associated and share some metabolic abnormalities, suggesting that they might also share some susceptibility genes. Thus, the SUR encoding gene is a plausible candidate for a primary pancreatic beta-cell defect and thus for hyperglycemia and weight gain. Through association and linkage studies, we have investigated the potential role of the SUR gene in families with NIDDM and in two independent sets of morbidly obese families. The exon 22 T-allele at codon 761 was more common in patients with NIDDM (7.7%) and morbid obesity (7.8%) than in control subjects (1.8%, P = 0.030 and P = 0.023, respectively). This variant was associated with morbid obesity (odds ratio 3.71, P = 0.017) and NIDDM (odds ratio 2.20, P = 0.04; association dependent on BMI). Although the frequencies for intron 24 variant were similar in all groups, morbidly obese patients homozygous for the c-allele had a more deleterious form of obesity. Sib-pair linkage studies with NIDDM in French Caucasian families gave no evidence for linkage to the SUR locus. However, in one set of the obese families, we found an indication for linkage with a SUR-linked microsatellite marker (D11S419, P = 0.0032). We conclude that in Caucasians, the SUR locus may contribute to the genetic susceptibility to NIDDM and obesity.

Recommendations regarding obesity surgery.

In France, 1,000 obese persons per month undergo a bariatric operation. Obesity surgery requires coordination and monitoring of aftercare. The French public health-care insurer asked the medical associations involved in obesity management to provide guidelines for obesity surgery. The recommendations were developed by the national associations of Obesity, Nutrition and Diabetes: the Association Française d'Etudes et de Recherches sur l'Obésité (AFERO), member of the EASO and IASO; the Association de Langue Française pour l'Etude du Diabète et des Maladies Métaboliques (ALFEDIAM); the Société Française de Nutrition (SFN); and the Société Française de Chirurgie de l'Obésité (SOFCO). This article presents the short version of the guidelines.

Chewing and swallowing as indices of the stimulation to eat during meals in humans: effects revealed by the edogram method and video recordings.

Patterns of chewing and swallowing were recorded during standardized meals in humans. Cocktail size (3 cm2) open sandwiches were served in one of five different flavors. An oscillographic recording of chewing and swallowing showed that chewing activity varied with the palatability and variety of foods. Chewing time was shorter and fewer chews were observed as palatability increased. Swallowing did not change as a function of stimulus flavor. Pause duration between two successive food pieces became shorter as palatability increased. The effects of sensory factors were most evident at the beginning of meals and decreased until the end of meals. A later study which compared eating parameters in sandwich, semi-solid, and traditional French meals (different courses ingested in succession: appetizer, main course, cheese and dessert), as assessed from video recordings, found that different microstructure parameters responded to palatability manipulation in different meal types. Strength of mastication and prandial drinking might be other important parameters to look at in order to understand the motivation to eat and its fluctuations during the meal.

Contraception in hypertensive women using a vaginal ring delivering estradiol and levonorgestrel.

Contraception with a vaginal ring (CVR) that delivers estradiol and levonorgestrel was used during a mean of 15.6 menstrual cycles in 12 hypertensive women. Blood pressure (BP) was measured 5 times on each visit during 2 pretreatment control cycles; during the 1st, 2nd, 4th, 6th, and from the 9th to 12th cycles of CVR use; and again after a 1-month recovery period. No significant change in BP occurred during CVR use in any of the subjects. Plasma renin substrate and antithrombin III activity did not vary significantly, which suggests the utility of administering natural estradiol via the vagina, thus avoiding the first pass effect that occurs with oral contraceptives. Significant decreases in plasma sex hormone-binding globulin, cholesterol, high density lipoprotein cholesterol, phospholipids, and triglycerides occurred, indicating an androgenic effect of levonorgestrel. We conclude that the CVR is a method of contraception that does not elevate BP in hypertensive women.

PIP: Contraception with a vaginal ring (CVR) that delivers estradiol and levonorgestrel was used during a mean of 15.6 menstrual cycles in 12 hypertensive women. Blood pressure (BP) was measured 5 times on each visit during 2 pretreatment control cycles; during the 1st, 2nd, 4th, 6th, and from cycles 9-12 of CVR use; and again after a 1-month recovery period. No significant change in BP occurred during CVR use in any of the subjects. Plasma renin substrate and antithrombin III activity did not vary significantly, which suggests the utility of administering natural estradiol via the vagina, thus avoiding the 1st pass effect that occurs with oral contraceptives. Significant decreases in plasma sex hormone-binding globulin, cholesterol, high density lipoprotein cholesterol, phospholipids, and triglycerides occurred, indicating an androgenic effect of levonorgestrel. The authors conclude that the CVR is a method of contraception which does not elevate BP in hypertensive women.

Biological effects of estradiol-17 beta in postmenopausal women: oral versus percutaneous administration.

To determine whether the route of administration or the type of estrogen used in estrogen replacement therapy (ERT) is more important in avoiding effects on hepatic function, 24 postmenopausal women were studied before and at the end of 2 months of oral or percutaneous administration of the same estrogen, estradiol-17 beta (E2). The treatments studied were oral micronized E2, 2 mg/day (9 women); oral E2 valerate, 2 mg/day (5 women), and percutaneous E2, 3 mg/day (10 women). Specific plasma biological and biochemical markers of estrogenic action were evaluated, namely, E2, estrone (E1), LH, FSH, sex steroid binding protein (SBP), renin substrate, antithrombin activity, and lipoproteins (high density lipoprotein cholesterol, low density lipoprotein cholesterol, very low density lipoprotein triglycerides). Both oral and percutaneous administration of E2 increased plasma E2 levels up to midfollicular values and decreased LH and FSH levels into the same range. Oral administration of E2 led to substantial increases in plasma E1, SBP, renin substrate, and VLDL levels, whereas AT decreased significantly. Percutaneous administration of E2 led to a physiological plasma E1/E2 ratio and did not induce any change in hepatic proteins. These data suggest that the route of administration of E2 determines the biochemical response to ERT in postmenopausal women. SBP is the most sensitive marker of the liver action of estrogen, and triglycerides also are simple and useful markers for this effect. Percutaneous E2 therapy is an effective method of ERT, and has no measurable effects on hepatic markers of estrogen action.

Mutation screening of the urocortin gene: identification of new single nucleotide polymorphisms and association studies with obesity in French Caucasians.

A linkage between obesity-related phenotypes and the 2p21-23 locus has been reported previously. The urocortin (UCN) gene resides at this interval, and its protein decreases appetite behavior, suggesting that UCN may be a candidate gene for susceptibility to obesity. We localized the UCN gene by radiation hybrid mapping, and the surrounding markers were genotyped in a collection of French families. Evidence for linkage was shown between the marker D2S165 and leptin levels (LOD score, 1.34; P = 0.006) and between D2S2247 and the z-score of body mass index (LOD score, 1.829; P = 0.0019). The gene was screened for SNPs in 96 obese patients. Four new variants were established. Two single nucleotide polymorphisms were located in the promoter (-535 A-->G, -286 G-->A), one in intron 1 (+31 C-->G), and one in the 3'-untranslated region (+34 C-->T). Association studies in cohorts of 722 unrelated obese and 381 control subjects and transmission disequilibrium tests, performed for the two frequent promoter polymorphisms, in 120 families (894 individuals) showed that no association was present between these variants and obesity, obesity-related phenotypes, and diabetes. Thus, our analyses of the genetic variations of the UCN gene suggest that, at least in French Caucasians, they do not represent a major cause of obesity.

Clinical studies with d-fenfluramine.

d-Fenfluramine (dF) (15 mg twice daily) has been studied in controlled trials in human obesity and has been shown to increase adherence to diet, to enhance its efficacy, and most importantly, to prevent weight regain when continued over 1 y. Few side effects, mostly transient, have been observed. A long-term use of dF in the management of some obese patients could be foreseen. Additionally, evidence that dF improves eating symptoms and dysphoric impairments in obese cravers, premenstrual syndrome, seasonal affective disorder, and smoking withdrawal syndrome has been presented.

Nutritional induction of adipose tissue lipoprotein lipase in obese subjects.

Other workers have reported increased adipose tissue lipoprotein lipase after a weight loss in obese subjects and have suspected that this enzyme is a primary factor of pathophysiological significance. In order to determine whether this effect was the consequence of refeeding rather than weight loss, six obese females were included in a controlled study. Adipose tissue lipoprotein lipase was measured before weight loss, at the end of 30 days on a diet of 800 kcal/day (mean weight loss 8.7%), and four times during the 8 days after the initiation of refeeding a 1500 kcal/day mixed diet to insure weight stability. Adipose tissue lipoprotein lipase decreased by 77% by the end of the weight loss, and an average 2-fold increase (24.2 +/- 2.7 mean +/- sem versus/11.1 +/- 2.3 mU/10(6) cells, p less than 0.01) was shown as early as 2 days after refeeding. Peak values after refeeding did not surpass predieting values. Changes during restriction and peak postrefeeding values were both positively correlated to baseline values. It can be concluded that the previously shown increase in lipoprotein lipase during weight stability after a weight loss is likely to be a secondary effect of partial refeeding; the individual sensitivity of adipose tissue lipoprotein lipase to nutritional induction could be of critical importance.

Relationships between 3-y longitudinal changes in body mass index, waist-to-hip ratio, and metabolic variables in an active French female population.

Three-year longitudinal changes in body mass index (BMI), waist-to-hip ratio (WHR), and metabolic variables were examined in 209 active French women. For the entire group, a weak but significant positive association was found between change in BMI and change in WHR. However, analysis of covariance according to the degree of abdominal fat distribution showed a heterogeneity of this association that was confined to women with abdominal fat distribution. Changes in BMI were positively associated with changes in total cholesterol (P less than 0.05), triglycerides (P less than 0.10), and blood pressure (P less than 0.001), whereas changes in WHR were associated with changes in triglycerides (P less than 0.05) and diastolic blood pressure (P less than 0.10). These longitudinal results suggest that a more favorable body-fat pattern and metabolic profile might be achieved by reducing weight, or at least by preventing weight gain, particularly in women with high abdominal-fat distribution.

How the degree of unsaturation of dietary fatty acids influences the glucose and insulin responses to different carbohydrates in mixed meals.

The association of fats with carbohydrates results in a lower glucose response but the influence of the nature of the dietary fatty acids has not been investigated clearly. We examined the effect of the degree of dietary fatty acid unsaturation on the postprandial glucose and insulin responses to a mixed meal. Eight young normolipidemic men consumed four different meals in random order. The meals differed in the nature of their oils and in the type of their main carbohydrates. The meals contained either a high ratio of monounsaturated to polyunsaturated n-6 fatty acids (MUFA) or a low ratio (PUFA) and either potatoes or parboiled rice. Proteins and saturated and polyunsaturated n-3 fatty acid contents were not different among meals. Blood samples were collected every 30 min for 3 h after the test meal. The glucose response was significantly lower 30 min after the parboiled rice-PUFA meal than after parboiled rice-MUFA or potato-MUFA (P < 0.05) meals. The insulin response was lower after parboiled rice-PUFA than after potato-MUFA (P < 0.05) meals. Similarly, an effect of fat appeared after 30 min. Glucose responses (F = 1.4, P < 0.01) and insulin responses (F = 5.3, P < 0.05) to both carbohydrates were significantly lower with dietary PUFA compared with dietary MUFA. In conclusion, the degree of dietary fatty acid unsaturation (18:1 compared with 18:2) may influence the glucose and insulin responses to mixed meals.

Satiating effect of proteins in healthy subjects: a comparison of egg albumin, casein, gelatin, soy protein, pea protein, and wheat gluten.

The influence of six dietary protein types (egg albumin, casein, gelatin, soy protein, pea protein, and wheat gluten) on satiety and food intake was investigated. Twelve healthy subjects ingested six protein-manipulated lunches (approximately 5.2 MJ, 22% of energy as protein) according to a within-subjects design. Test meals were controlled for energy, macronutrients, fiber, and palatability. Nearly 65% of total protein varied between sessions. After lunch, satiety was assessed for 8 h and energy and macronutrients intakes were measured for 24 h. Blood was collected for determination of postprandial plasma glucose and insulin responses. Results showed no effect of the type of protein on satiety, on 24-h energy or macronutrient intakes, or on postprandial plasma glucose and insulin concentrations. These findings differ in part from those obtained previously in humans, which suggested that proteins may be differentiated in terms of their satiating capacities. We conclude that varying the protein source in a mixed meal does not affect food behavior in healthy humans, probably because coingestion of carbohydrate and fat with protein buffers the kinetics of the physiologic mechanisms implicated in postprandial satiety after a protein load.

Evidence for a cholesteryl ester donor activity of LDL particles during alimentary lipemia in normolipidemic subjects.

Postprandial hypertriglyceridemia represents an independent risk factor for coronary artery disease. In the postprandial state, elevated levels of triglyceride-rich lipoproteins (TRL) are minor acceptors of HDL-cholesteryl ester (CE) transferred by CETP in normolipidemic subjects: indeed, LDL particles represent the major CE acceptors. In order to evaluate further the potential atherogenicity of lipoprotein particles characteristic of the postprandial phase in normolipidemic subjects, we determined the quantitative and qualitative features of apoB- and apoAI-containing lipoproteins over an 8-h period following consumption of a mixed meal. During postprandial lipemia, we observed a significant decrease (-12%) in plasma AI concentration (138+/-4 and 156+/-4 mg/dl, at 3 h and baseline, respectively, P<0.005). Concomitantly, a progressive increase (+13%) was detected in HDL2 concentrations (138+/-7 mg/dl at 4 h vs. 122+/-12 mg/dl at baseline, P<0.005), as well as a significant reduction (-9%) in HDL3 levels (137+/-6 mg/dl at 3 h vs. 150+/-4 mg/dl at baseline; P<0.05). Additionally, plasma LDL was reduced by 5% (247+/-12 mg/dl at 3 h vs. 260+/-15 mg/dl at baseline; P<0.05) 3 h following meal intake. Moreover, a significant reduction (-10%) occurred in the CE/TG ratio in LDL at 2 h postprandially (8+/-2 at 2 h vs. 9+/-3 at baseline; P<0.005). These changes reflected an increment (17+/-3 mg/dl at 3 h vs. 15+/-4 mg/dl at baseline; P<0.05) in LDL triglyceride concentrations. Despite the high CE acceptor capacity of LDL particles, no measurable increase in their CE content was detected during the postprandial phase. We demonstrated that CE accepted by LDL particles from HDL are secondarily transferred to chylomicrons by CETP. As chylomicrons displayed a 260-fold lower CE/TG ratio than LDL (0.03:1 and 7.8:1 in chylomicrons and LDL, respectively), CE-rich LDL may act to donate CE to chylomicrons. In conclusion, our data indicate that the presence of elevated levels of chylomicrons induces LDL to act as a secondary donor of CE during the postprandial phase.

Hormonal influences on the relationships between body fatness, body fat distribution, lipids, lipoproteins, glucose and blood pressure in French working women.

The independent associations between overall obesity, body fat distribution, lipids, lipoproteins, glucose, blood pressure and some hormonal factors (sex hormone-binding globulin (SHBG), corticosteroid binding globulin (CBG) and fasting insulin) were cross-sectionally examined in 205 French working women. After adjustment for age, overall adiposity assessed by body mass index (BMI) was significantly associated with most metabolic parameters, whereas regional adiposity assessed by the waist-hip ratio (WHR) was significantly associated only with triglyceride, systolic and diastolic blood pressure. Blood pressure, glucose but not triglyceride, were also negatively and significantly correlated with SHBG and positively with fasting insulin. Negative independent associations were found between SHBG and both BMI and WHR, whereas CBG was positively associated only with WHR. Fasting insulin was no longer related to WHR after adjustment for BMI. After controlling for the effect of SHBG or insulin, the associations between triglyceride, blood pressure and both BMI and WHR were not substantially modified. After adjustment for BMI and WHR, fasting insulin was independently associated with both HDL cholesterol and diastolic blood pressure. In conclusion, in these French women, hormonal factors under study appeared to have little influence on the relationships between body fatness, body fat distribution, metabolic variables and blood pressure.

Skinfold and body circumferences as measures of body fat patterning in a French female active population: relationships with the metabolic risk profile.

Body mass index (BMI), various anthropometric indices of abdominal fat distribution and some metabolic variables (blood lipids, fasting glucose, blood pressure) were measured in 408 French-born women from an occupational population who volunteered for the study. The aim of the study was to determine the best index for describing the relationships between the body fat pattern and the metabolic risk profile. The four age-adjusted circumference ratios (waist/hip, waist/thigh, xiphoid/hip, xiphoid/thigh) showed similar associations with the metabolic variables whereas the three age-adjusted skinfold ratios (epigastric/thigh, mesogastric/thigh, hypogastric/thigh) tended to be more weakly associated with the metabolic variables, particularly with apolipoprotein (Apo) B and fasting glucose. Multiple regression analyses showed that age-adjusted BMI was significantly related to high density lipoprotein (HDL) cholesterol, low density lipoprotein (LDL) cholesterol, Apo A and Apo B, and blood pressure, independently of abdominal fat distribution. After controlling for the effects of BMI, the waist/thigh ratio remained significantly associated to triglyceride, Apo B, fasting glucose, and systolic blood pressure, whereas the waist/hip ratio and the mesogastric/thigh skinfold ratio were significantly related only to triglyceride and systolic blood pressure independently of BMI. With the exception of triglyceride and fasting glucose, the degree of association between the metabolic variables and the abdominal fat distribution tended to be weaker than that observed with the BMI. These results emphasize the importance of the global corpulence in the levels of metabolic variables. However, all indices of abdominal fat distribution were, to varying degrees, independently associated with an unfavorable metabolic profile. Among them, the waist/thigh circumference ratio seems to be a useful indicator of the body fat pattern in women.