Facilitated long chain fatty acid uptake by adipocytes remains upregulated relative to BMI for more than a year after major bariatric surgical weight loss.

OBJECTIVE: This study examined whether changes in adipocyte long chain fatty acid (LCFA) uptake kinetics explain the weight regain increasingly observed following bariatric surgery.
METHODS: Three groups (10 patients each) were studied: patients without obesity (NO: BMI 24.2 ± 2.3 kg m(-2) ); patients with obesity (O: BMI 49.8 ± 11.9); and patients classified as super-obese (SO: BMI 62.6 ± 2.8). NO patients underwent omental and subcutaneous fat biopsies during clinically indicated abdominal surgeries; O were biopsied during bariatric surgery, and SO during both a sleeve gastrectomy and at another bariatric operation 16 ± 2 months later, after losing 113 ± 13 lbs. Adipocyte sizes and [(3) H]-LCFA uptake kinetics were determined in all biopsies.
RESULTS: Vmax for facilitated LCFA uptake by omental adipocytes increased exponentially from 5.1 ± 0.95 to 21.3 ± 3.20 to 68.7 ± 9.45 pmol/sec/50,000 cells in NO, O, and SO patients, respectively, correlating with BMI (r = 0.99, P < 0.001). Subcutaneous results were virtually identical. By the second operation, the mean BMI (SO patients) fell significantly (P < 0.01) to 44.4 ± 2.4 kg m(-2) , similar to the O group. However, Vmax (40.6 ± 11.5) in this weight-reduced group remained ~2X that predicted from the BMI:Vmax regression among NO, O, and SO patients.
CONCLUSIONS: Facilitated adipocyte LCFA uptake remains significantly upregulated ≥1 year after bariatric surgery, possibly contributing to weight regain.

Introduction

Obesity is the accumulation of excess fat, principally triglycerides (TG), in adipocyte depots throughout the body. Excessive TG, typically within discrete droplets, also accumulates in the liver and heart, where they are responsible for clinical consequences such as nonalcoholic fatty liver disease [1] and obesity cardiomyopathy [2]. In the late 1980s, skeptical of the concept that long chain fatty acids (LCFA) entered cells exclusively by diffusion, we postulated that the principal uptake process would prove to be regulatable, facilitated transport, and used studies of cellular LCFA uptake kinetics in rodents to prove that [3-7]. We identified the first LCFA transporter [8-11], showed that regulation of LCFA uptake in adipocytes was a control point for adiposity [12-14], and that up-regulation of facilitated LCFA uptake in hepatocytes [1] & cardiomyocytes [2,15] was a key element in pathogenesis of obesity- and EtOH-associated hepatic steatosis & cardiomyopathies in rodents.

Translational studies confirmed key findings in man, especially up-regulation of adipocyte LCFA uptake in patients with obesity[16]. The present study extends these observations to patients classified as super-obese participating in a 2-stage bariatric surgical protocol beginning with a sleeve gastrectomy [17]. After major weight loss over the first post-operative year, weight typically stabilized during year 2, leading to a second procedure, usually a biliopancreatic diversion with duodenal switch (BPD-DS), in patients requiring further weight loss. Fat biopsies from these procedures facilitated studies of the effects of weight loss on multiple aspects of adipocyte biology.

Study Hypothesis

The rate (Vmax) of facilitated (saturable) LCFA uptake into human adipocytes is highly correlated with BMI and % body fat, increasing in obesity and decreasing with weight loss.

Study Aims

(1) to compare the Vmax for LCFA uptake with BMI in four patient groups: [1] patients who were not obese (NO), [2] patients with obesity (O), [3] patients classified as super-obese (SO), and (2) to compare Vmax with BMI in SO patients both at the time of an initial sleeve gastrectomy and, at a second surgical procedure after major weight loss, when they were classified as [4] super-obese reduced (SOr).

Methods

Patients and Protocol

Patients classified as Super-Obese (SO)

After stabilizing their weight for ca. 2 months following recommended dietary modifications, SO patients (BMI >50 kg/m2) at Weil-Cornell & Columbia University Medical Centers underwent an initial laparoscopic sleeve gastrectomy. When their weight subsequently stabilized, after significant weight loss often accompanied by remission of diabetes and/or hypertension, 10 returned for a second operation [17].

We received omental and subcutaneous fat biopsies and a venous blood sample at both surgical procedures in order to examine the effects of surgical weight loss on aspects of adipocyte biology and adipocyte LCFA uptake. Power calculations indicated that at least 10 patients would have to complete both operations to reach our desired end-points with appropriate statistical assurance. Accordingly, the study remained open to enrollment until 10 patients, now designated as SOr, had completed second stage surgery. By chance, those 10 patients consisted of 5 men and 5 women. As anticipated by prior experience, a total of 35 SO patients were enrolled to meet the study needs.

Patients who were Not-Obese (NO) (BMI <30)

NO patients consented to donate a venous blood sample and omental and subcutaneous fat biopsies during a clinically indicated laparoscopic abdominal surgical procedure. Of 18 NO patients studied, 5 men and 5 women were selected to age-match this cohort as closely as possible with the SO patients. Of the 10, 7 had undergone donor nephrectomy, 2 laparoscopic cholecystectomy, and 1 an inguinal herniorraphy.

Patients with Obesity (O) (BMI > 35)

Of 34 O male and 49 O female bariatric surgery patients studied by our laboratory, 5 men and 5 women were similarly chosen on the basis of age for inclusion in this analysis. The minimum BMI of 35 in this group reflects the lower limits of BMI currently acceptable for bariatric surgery.

Materials

[3H]-oleic acid (OA) was from NEN Life Science Products (Boston, MA, USA), type I collagenase from Sigma (St Louis, MO, USA), and fatty acid-free bovine serum albumin (BSA) from Boehringer Mannheim (Indianapolis, IN). Circulating Spexin was assayed with a competitive EIA kit and leptin by an antigen capture ELISA kit from Phoenix Pharmaceuticals (Burlingame, CA).

Body Fat

Percent body fat was measured in SO & SOr patients using Tanita scales equipped with Bioelectrical Impedance technology.

Isolation of adipocytes

Biopsies of 5 - 10 g were obtained from all enrolled patients from omental & anterior abdominal wall subcutaneous fat depots at each operation. Approximately 1/3 of each biopsy was frozen at −80°C in RNALater for subsequent qRT-PCR gene expression & biochemical studies. Adipocyte single cell suspensions meeting established viability criteria were prepared with collagenase from the remainder of each biopsy and counted as described[10, 16, 18, 19].

Cell size studies

Diameter distributions in each adipocyte preparation were determined as reported by digital analysis of suspended cells using a Nikon Eclipse 80i microscope and Nikon Digital DXM 1200C camera. Digital images were analyzed using Nikon NIS-Elements (NE) Br software, generating mean diameters (with distribution), in micrometers (μ), for each preparation. The mean cell surface area and cell volume, in μ2 and pl, respectively, were computed from the diameters [20, 21].

Adipocyte LCFA uptake studies

Aliquots from each cell preparation were incubated at 37°C in Dulbecco's Modified Eagle's Medium (DMEM) containing 500 μM BSA and one of five different concentrations of OA, such that the OA:BSA molar ratio (ν) was 0.25, 0.5, 1.0, 1.5, or 2.0:1 [3]. The initial velocity (V0) of cellular OA uptake from each test solution was determined by a rapid filtration technique from four samples obtained in triplicate over the initial 30 s of incubation, during which uptake was a linear function of time [3, 5, 7, 10, 14, 16, 22].

Computations and data fitting

The unbound oleate concentration ([OAu]) in each test solution was calculated from ν [23], using the LCFA:BSA binding constants of Spector et al [24]. Measurements of initial OA uptake velocity (V0) at values of ν from 0.25-2.0 were fitted to the sum of saturable & nonsaturable functions of the corresponding [OAu][7] according to the equation: UT([OAu]) = Vmax • [OAu] /(Km + [OAu]) + k • [OAu]. UT([OAu]) is the experimental measurement of uptake, in pmol/sec/50,000 cells, at the stated concentration of unbound OA ([OAu]); Vmax & Km are the maximal velocity of saturable OA uptake and the value of [OAu] at one-half the maximal uptake velocity; k is the rate constant for nonsaturable uptake.

Data fitting used the SAAM II version of the Simulation, Analysis & Modeling (SAAM) program of Berman and Weiss [25, 26] to compute for each data set values of Vmax (pmol/sec/50,000 cells), Km (nM), and k (ml/sec/50,000 cells), and their variances & covariances. Prior studies showed that under the conditions employed, V0 and derived parameters such as Vmax are measures of actual transmembrane transport [3, 5, 27] Further studies demonstrated that an increase in Vmax preceded an increase in adipocyte size early in the development of obesity [13], and a decrease in Vmax preceded a reduction in adipocyte size and body weight during leptin-induced weight loss [14], showing that changes in Vmax do not simply reflect changes in cell size.

Statistical Methods

Relationships between parameters were assessed by both linear and nonlinear correlations [28]. For group comparisons, results are expressed as mean ± SE, with n = 10 per group. Each of the experimental groups was compared to the control group with two-tailed Student's t-tests. The other groups were also compared with each other by one way ANOVA as previously described [15]. In addition, the effects on changes in LCFA uptake rates in response to weight loss, of age, gender, ethnicity, baseline weight, % body fat, metabolic status (as reflected in e.g. HbA1c and cholesterol), and the presence of specific co-morbidities or medication use, were explored by effect adjustments in the ANOVA. In all statistical testing, significance was set at p≤0.05.

Spexin and leptin gene expression and serum assays

Circulating Spexin was measured by competitive enzyme immunoassay (EIA) & leptin by antigen capture ELISA using Phoenix Pharmaceuticals kits (Burlingame, CA) [29]. Sera were diluted 1/20 in assay buffer, and quantified by comparison to within-assay standard curves according to the manufacturer's instructions.

Results

Patients

Demographic and clinical laboratory data for the 10 participants in each of the NO, O, and SO groups are summarized in , as are analogous data for the group designated as SOr, which were obtained from SO patients at the time of their second bariatric procedure. Mean ages, initial BMIs and clinical and laboratory data for the 10 SO patients who completed a second operation are very similar to corresponding data from all 35 SO patients enrolled in the study. Overall, the NO, O, and SO patient groups were similar in age (). O and SO weighed more than NO patients and had higher BMIs (p<0.001). While high density lipoprotein (HDL) values were lower and triglycerides (TG) higher in the O and SO patients than in the NO controls (), there were no significant increases in glucose or cholesterol in these two groups of obese patients, possibly reflecting ongoing treatment for hyperglycemia and/or hypercholesterolemia. Albumin was marginally reduced and aspartate aminotransferase (AST) and/or alanine aminotransferase (ALT) marginally increased in the SO and O groups (). Overall, the abnormalities in aminotransferases were on the mild end of the spectrum observed in larger populations of super-obese patients [30]. Gender differences were observed for several clinical and laboratory parameters summarized in . However, as there were only 5 patients of each gender per group, these differences were not analyzed further.

Additional data from the 10 individual SO/SOr patients that were, inter alia, the basis for our ANOVA effects adjustment testing are presented in . Among these 10 SO patients, 8 were Caucasian-non-Hispanic, 1 Caucasian-Hispanic, and 1 self-described as a Caucasian/ Hispanic/ African American (). Initially 8 of the 10 SO patients had hypertension and 6 had diabetes. SO patients had a mean of 3 comorbidities each; 1 patient had only steatohepatitis, and only 1 patient had none.

Adipocyte Dimensions

Diameters of both omental and subcutaneous adipocytes from the 30 NO, O & SO patients increased linearly with increasing BMI (Omental: r=0.55, p<0.01; Subcutaneous: r=0.46, p <0.05) (). Computed adipocyte surface areas and volumes were similarly correlated with BMI. When averaged by group, the calculated omental adipocyte surface area increased from 23.2±4.1 to 34.8±2.0 to 39.1±4.5 X 103 μ2 per cell and cell volume from 437±74 to 616±53 to 749±124 pl per cell in NO, O & SO patients. Subcutaneous adipocytes showed very similar trends.

Adipocyte LCFA Uptake Studies

The Vmax for facilitated LCFA uptake by omental adipocytes increased exponentially as a function of BMI across the 3 patient groups (). Mean values averaged 8.3±0.6, 20.9±1.4, and 68.7±9.6 pmol/sec/50,000 cells in the NO, O, and SO patients, respectively. Data from subcutaneous adipocytes were very similar (). The Vmax for omental and subcutaneous adipocytes in individual patients in each of the three patient groups were also very similar, and were linearly related (r= 0.96) (. Effects of surgical weight loss on Vmax () are examined below. Omental Vmax values in SO & O patients were highly correlated with BMI (p<<0.01) (), which in turn was highly correlated with Body Fat % (p<0.01) (). Results in subcutaneous adipocytes were virtually identical.

Effects of Post-Surgical Weight Loss

SOr participants who returned for a second operation 16.3±2.2 mos after their initial surgery had lost a mean of 113±13 lbs. As indicated by comparing SO with SOr results in , weight loss was associated with modest changes in clinical laboratory parameters, including glucose, lipids, and liver tests. Medications for control of T2DM, hyperlipidemia, or cardiovascular disorders including hypertension, prescribed by referring physicians, were also little changed ().

LCFA Uptake Kinetics After Weight Loss

SOr patients exhibited appreciable reductions in the Vmax for LCFA uptake in both omental and subcutaneous adipocytes when compared to values at initial surgery. At their 2nd surgery, BMIs had fallen from their initial 62.6±2.8 kg/m2 to 44.4±2.4 kg/m2, a value not significantly different from the 50.1±1.1 kg/m2 in the O group (). There were corresponding reductions in Body Fat % (). Both omental and subcutaneous adipocyte dimensions () fell into the NO range in most SOr patients following initial bariatric surgery, and were reduced compared to those of steady-state O adipocytes. However, Vmax's for facilitated adipocyte LCFA uptake (Omental: 42.1±6.4 pmol/sec/50,000 cells; subcutaneous: 37.7±6.2 pmol/sec/50,000 cells) remained significantly increased to ~2X that predicted for their BMI by the BMI:Vmax regression among NO, O, & SO patients (, 2X the value observed in the O patient group and ca. 5-fold compared to the NO range (, indicating persistent up-regulation of both omental and subcutaneous facilitated adipocyte LCFA uptake in SOr patients. As illustrated in , the omental Vmax fell appreciably in the transition from SO to SOr status in 9 of 10 individual patients, the exception being a patient who - unknown to his surgeon at the time - began to drink heavily in the period between his two bariatric surgical procedures.

In our model of LCFA transport [7, 13, 14, 16], Vmax’ is defined as a measure of the density of LCFA transporters per unit of cell surface area (Vmax’ = Vmax/Cell Surface Area) (pmol/sec/ μ2 X 10−8 cell surface) (). Vmax’, on average, increased with weight loss in omental SOr vs SO adipocytes, and were elevated ~5-fold compared to the NO range and 3.7-fold compared to that in Obese adipocytes (. By contrast, in a few SOr patients, Vmax’ decreased rather than increasing with weight loss (. Other than being a means of classifying adipocytes in terms of the response of their fatty acid uptake process to weight loss, the precise physiological implications of the nature of the change in Vmax’ require further exploration (see . By contrast to Vmax’, k’, a measure of non-saturable (passive) uptake per unit surface area, was not significantly changed in any group, consistent with its previously ascribed role as a measure of cell membrane permeability to passive LCFA diffusion [14, 16]. The means of the 40 values of k’ in the NO, O, SO, SOr groups were 0.0034±0.0005 X 10−8 & 0.0033±0.0004 X 10−8 ml/sec/μ2 cell surface area for omental and subcutaneous adipocytes, respectively.

Effects on Serum Leptin and Spexin Concentrations

Human serum leptin concentrations have a positive, non-linear correlation with BMI; in contrast, Spexin concentrations and BMI are negatively correlated [29]. These and others findings raise the possibility that leptin and Spexin are counter-acting regulators of adipocyte LCFA uptake [29]. In this study leptin concentrations in SO patients fell from 79.61 ± 22.05 to 37.72 ± 14.23 ng/mL (mean ± SE, p = 0.009) while serum Spexin increased from 1.65 ± 0.41 to 2.4 ± 0.36 ng/mL (p = 0.043) between their two bariatric surgeries, consistent with the previously identified trend [29].

Modulators of Changes in LCFA Uptake With Weight Loss

No significant effects of age, gender, ethnicity, baseline weight, metabolic status (e.g. HbA1c, cholesterol), presence of specific co-morbidities or medication use on the observed changes in LCFA uptake rates in response to weight loss were detected by the effect adjustments in the ANOVA. However, this conclusion must be qualified because of the small group sizes available for these analyses.

Discussion

The complexity of changes in adipose gene expression were not appreciated [31] and Spexin [32] and its association with weight regulation [29] had not been discovered when the project began in 2006. Our initial expectations were that adipocyte size and saturable LCFA uptake would fall in parallel during bariatric surgery-induced weight loss.

While omental adipocyte sizes () and the Vmax for LCFA () uptake both decreased as SO patients’ BMIs fell after initial bariatric surgery, they did not decrease in parallel. SOr cell sizes often fell into the normal range and were therefore small relative to BMI, but Vmax did not consistently fall even into the O range in these patients, whose BMIs remained in the 40s. Consequently, when expressed per unit surface area, Vmax’ of omental adipocytes actually increased 28%, from 3.6 ± 0.5 to 4.6 ± 0.9 pmol/sec/μ2 X 10−8, a value nearly 4-fold higher than the 1.2 ± 0.1 pmol/sec/μ2 X 10−8 typical of O patients (). These findings are graphed () and their implications discussed in detail in . Our cell size and Vmax’ data are consistent with recent models [33] which propose that small for body weight adipocytes are an important stimulus to weight regain via several complex pathways.

Since up-regulation of adipocyte LCFA transport is closely associated with obesity in man [16], its up-regulation predicts weight gain in multiple animal obesity models [1, 12, 13], & this study documents a significant correlation between LCFA uptake and BMI (), it is tempting to speculate that persistent upregulation of LCFA uptake in our SOr patients following bariatric surgical weight loss is a harbinger of weight regain. Weight regain has become an increasingly important issue in obesity management, after both dietary weight loss [34] and bariatric surgery [35]. The LABS consortium recently published the outcomes over 3 years post bariatric surgery in 2458 obese patients [36], of whom 1738 underwent Roux-en-Y gastric bypass (RYGB), 610 laparoscopic placement of an adjustable gastric band (LAGB), and 110 other procedures including 59 sleeve gastrectomies. RYGB and LAGB patients experienced most of their total weight loss in the first post-operative year. To evaluate weight patterns, 5 weight trajectory groups were identified for each procedure. The 5 RYGB trajectories all showed initial weight loss for 6 months after surgery, but by the third post-operative year, trajectories for all 5 groups demonstrated weight regain, accompanied by some recurrences of co-morbidites. Modest weight regain was reported during the second and third post-operative years among patients who had undergone sleeve gastrectomy or RYGB in another trial comparing intensive medical therapy alone vs intensive medical therapy plus bariatric surgery for treatment of diabetes [37]. The prevalence of weight regain in these and earlier studies[35] became increasingly evident by the middle of the second post-operative year, corresponding to when we noted persistent up-regulation of adipocyte LCFA uptake in our SOr patients.

Persistence of weight gain-promoting hormone patterns; increased insulin sensitivity, rates of glucose transport, and LPL activity; and a multiplicity of persistent metabolic abnormalities are factors believed to contribute to weight regain [33, 35, 38]. Given our earlier demonstration of the association between weight gain and up-regulation of adipocyte LCFA uptake, and of the present results, it is tempting to speculate that persistent up-regulation of adipose tissue LCFA uptake is, at the least, another potential mechanism contributing to weight regain after initial weight loss induced by bariatric surgery. However, to prove that, it will be necessary to study it, other potential causes of weight regain, and weight regain per se in the same cohort. This was impossible in the current study because our protocol mandated a second omental fat biopsy, which could only be obtained during a second bariatric surgical procedure. This second procedure led to further weight loss, averaging 50 ±13.5 lbs by a mean of 11 months post-operatively, making any tendency to more modest weight regain from the first surgery undetectable. However, the finding that adipocyte dimensions and LCFA uptake kinetics in subcutaneous adipocytes are virtually identical to those in omental fat is important, since subcutaneous fat can be obtained by aspiration during routine outpatient visits. Correlating serial aspiration biopsies of subcutaneous fat and simultaneous weight determinations after a single bariatric surgical procedure in each patient will provide a much stronger assessment of the relationship between LCFA kinetics and weight regain, indicate whether any observed up-regulation of adipocyte LCFA uptake persists with longer follow-up or whether it eventually normalizes, and the extent to which it occurs with all bariatric surgical procedures or is a unique consequence of sleeve gastrectomy. Adaptive responses lead the majority of patients who were previously obese, who then lost weight by dietary restriction, to later regain the lost weight [33, 34, 38]. The role for persistently up-regulated adipocyte LCFA uptake in that process is an open question.

Body weight and energy balance are principally regulated by integration of numerous signals, including concentrations of hormones released mainly from the gut and adipose tissues [39]. The details of these processes are still being elucidated, as is the extent to which the persistent up-regulation of LCFA uptake reflects abnormal hormonal patterns or a more complex pathogenesis.

Several large studies [35,36,40] show that bariatric surgery is the most effective current approach to short- and medium-term weight reduction and, often, remission of co-morbidities. However, its longer term efficacy and the role of weight regain in modulating its benefits are still uncertain. Since effective anti-obesity drugs developed in response to improved understanding of obesity pathophysiology are likely to become available in the future, appropriate therapeutic choices for optimal weight management will require improved understanding of the underlying physiology. Accordingly, evaluating the impact of persistently increased LCFA uptake & other mechanisms on weight regain and long-term obesity management should be actively pursued, and relevant processes, including increased adipocyte LCFA uptake, identified & studied in detail for their possible therapeutic benefits.

Table 1

Clinical and Biochemical Characteristics of the Four Patient Groups

	[1]Age	BMI	Glucose	Cholesterol	HDL	LDL	TG	Albumin	Bilirubin	AST	ALT	Alk Phos
	(yrs)	(kg/M[2])	(mg/dL)	(mg/dL)	(mg/dL)	(mg/dL)	(mg/dL)	(mg/dL)	(mg/dL)	(U/L)	(U/L)	(U/L)
NON OBESE
Male	44.8 ± 7.0	24.1 ± 1.3	110.3 ± 17.3	179.3 ± 13.5	47.3 ± 5.7	96.8 ± 7.28	109.5 ± 30.9	4.6 ± 0.1	0.5 ± 0.1	17.8 ± 2.1	23.8 ± 6.9	68.8 ± 4.3
Female	40.8 ± 4.8	23.3 ± 0.5	79.3 ± 1.6	178.3 ± 19.1	59.3 ± 5.4	106.8 ± 16.89	61.5 ± 11.1	4.3 ± 0.1	0.5 ± 0.1	18.0 ± 2.9	15.5 ± 3.4	79.0 ± 18.7
M vs F	[2]NS	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS
Total	42.8 ± 4.0	23.7 ± 0.7	94.8 ± 9.9	178.8 ± 10.6	53.3 ± 4.3	101.8 ± 8.72	85.5 ± 17.7	4.5 ± 0.1	0.5 ± 0.1	17.9 ± 1.6	20.1 ± 3.8	73.3 ± 8.1
OBESE
Male	47.6 ± 1.6	49.5 ± 5.5	90.4 ± 6.1	107.6 ± 7.1	33.8 ± 3.4	84.8 ± 5.8	144.6 ± 24.7	4.1 ± 0.1	0.8 ± 0.1	25.2 ± 3.4	40.2 ± 7.9	68.8 ± 5.4
Female	49.6 ± 3.2	50.5 ± 1.5	94 ± 4.0	216.5 ± 17.9	46.0 ± 1.8	145.8 ± 17.8	123.8 ± 18.5	4.2 ± 0.2	0.5 ± 0.1	20.0 ± 3.4	19.4 ± 3.6	72.6 ± 8.3
M vs F	NS	NS	NS	p<0.001	p<0.01	p<0.01	NS	NS	p<0.025	NS	p<0.025	NS
Total vs NO	48.6 ± 1.7NS	50.0 ± 2.7p<0.001	92.2 ± 3.5NS	156 ± 15.9NS	39.2 ± 2.9p<0.01	111.9 ± 14.3NS	136.3 ± 15.1p<0.05	4.2 ± 0.1p<0.05	0.7 ± 0.1NS	22.8 ± 2.5NS	30.8 ± 5.0NS	70.5 ± 4.7NS
SUPER OBESE
Male	51.6 ± 2.9	61.5 ± 6.0	88.8 ± 3.7	175.2±19.4	41 ± 3.3	105.4 ± 12.8	142.8 ± 45.6	3.9 ± 0.1	0.7 ± 0.2	26.6 ± 4.9	32.4 ± 8.1	66 ± 12.8
Female	38.5 ± 3.9	63.8 ± 2.1	103.5 ± 14.6	185.8 ± 17.4	43.3 ± 5.6	122.8 ± 14.9	142.3 ± 43.7	3.5 ± 0.2	0.5 ± 0.1	26 ± 4.3	28.4 ± 5.2	70.2 ± 6.8
M vs F	p<0.025	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS	NS
Total vs NO	44.5 ± 3.1NS	62.8 ± 2.9p<0.001	95.3 ± 5.6NS	180.5 ± 11.7NS	42.0 ± 2.4p<0.05	113.1 ± 8.5NS	142.6 ± 26.4p<0.025	3.7 ± 0.1p<0.001	0.6 ± 0.1NS	26.3 ± 2.2p<0.005	30.4 ± 4.3p<0.025	68.1 ± 6.8NS
SUPER OBESE Reduced
Male	52.8 ± 2.7	42.8 ± 4.0	73.2 ± 5.9	165.4 ± 25.7	47.2 ± 4.7	96.6 ± 17.4	112.2 ± 39.1	4.9 ± 0.2	0.6 ± 0.1	24.6 ± 2.9	21.2 ± 2.3	61.0 ± 10.4
Female	39.2 ± 4.3	46.7 ± 4.5	89.4 ± 8.0	217.3 ± 18.9	49.0 ± 12.5	143.0 ± 17.2	127.3 ± 40.1	3.7 ± 0.2	0.6 ± 0.1	18.4 ± 3.7	12.6 ± 1.3	66.0 ± 12.8
M vs F	p<0.05	NS	NS	NS	NS	p<0.05	NS	NS	NS	NS	p<0.005	NS
Total vs NO	46.0 ± 3.3NS	45.2 ± 2.7p<0.001	81.3 ± 5.3NS	184.9 ± 18.6NS	47.9 ± 4.5NS	114.0 ± 14.3NS	117.9 ± 25.3NS	3.9 ± 0.1p<0.001	0.6 ± 0.1NS	21.5 ± 2.4NS	16.9 ± 1.9NS	63.5 ± 8.5NS

All data Mean ± SE

NS = not significantly different (p>0.05)