Introduction
Reports of fat redistribution and metabolic disturbances in HIV-infected patients receiving potent antiretroviral therapy began to appear in 1998 and together these changes have been referred to as a 'lipodystrophy syndrome'. The majority of cases of fat redistribution and metabolic disturbances have been described in patients treated with protease inhibitor (PI)-based antiretroviral therapy [1], but subsequent cases have been described in individuals never treated with these agents [2,3].
Changes in fat distribution in HIV-infected patients include regional lipohypertrophy and lipoatrophy [4]. Increased fat accumulation is seen in the visceral depot and less commonly in the dorsocervical fat pad and female breast. By contrast, loss of fat occurs in the subcutaneous depots of the extremities, buttocks and face. Manifestations of both lipohypertrophy and lipoatrophy are present in many patients.
Metabolic changes include insulin resistance and an atherogenic lipid profile characterized by elevated low density lipoprotein cholesterol (LDL-C) and triglyceride levels and decreased high density lipoprotein chloesterol (HDL-C) levels [1]. PI may play an independent role in any one or in all of these metabolic changes. On the other hand, the metabolic disturbances may be a consequence of changes in body fat distribution. This distinction is important because different pathogenic mechanisms for the metabolic disturbances associated with PI therapy are implied.
In the general population, cross-sectional and longitudinal studies show that high levels of abdominal adipose tissue are associated with insulin resistance, glucose intolerance and a dyslipidemia characterized by elevated triglyceride levels and low HDL-C levels [5-9]. An association between central adiposity and insulin resistance has also been demonstrated in the HIV lipodystrophy syndrome without an independent effect of PI therapy on this metabolic parameter [10]. However, short-term PI therapy causes insulin resistance independent of body fat changes in healthy HIV-uninfected males [11]. Short-term interruption of PI-based antiretroviral therapy improves triglyceride and LDL-C levels, but not insulin or glucose levels [12]. To investigate whether insulin resistance and other metabolic changes are related to changes in body composition in HIV-infected patients or to PI therapy per se, we studied PI-treated patients with and without lipodystrophy. HIV-infected subjects who had not received PI therapy and who had no change in fat distribution served as controls. Because increased resting energy expenditure (REE) has been described in other forms of lipodystrophy, this was measured to determine if hypermetabolism is a feature of the lipodystrophy syndrome associated with HIV infection.
Methods
HIV-infected subjects were recruited from local HIV primary care practices. Subjects gave written informed consent under a protocol approved by the Institutional Review Board at the University of Colorado Health Sciences Center. PI-treated subjects were included in the group with lipodystrophy (PI-LD), if the subject, the subject's primary care provider and the primary investigator reported accumulation of central fat in addition to loss of fat from at least one depot, and the waist-hip ratio (WHR) was > 0.95 in men and > 0.88 in women. These ratios were chosen because they are widely accepted as valid predictors of central adiposity in the general population. PI-treated and PI-naive subjects were included in the groups without lipodystrophy (PI-controls and non-PI-controls, respectively) if the subject, the subject's primary care provider and the primary investigator agreed that the patient showed no signs of lipohypertrophy or lipoatrophy and the WHR was < 0.95 in men and < 0.88 in women. Subjects were excluded if they had a documented history of hyperlipidemia prior to PI therapy, were treated with high-dose glucocorticoid therapy in the past year or had an active opportunistic infection or malignancy. All subjects were studied during inpatient admissions to the General Clinical Research Center at the University of Colorado Health Sciences Center.
Blood samples for analysis of serum lipoproteins, insulin, glucose, plasminogen activator inhibitor-1 (PAI-1) activity, fibrinogen, CD4 cell counts and viral load measurements were collected after a 12 h overnight fast. PAI-1 activity was determined using the Stachrom-PAI-1 kit (Diagnostica Stago, Asnieres-sur-Seine, France) and performed on an ACL 3000 plus analyzer (Instrumentation Laboratories, Lexington Massachusetts, USA). Fibrinogen was measured by mechanical clot detection using Fibrinogen 5 (Diagnostica Stago) and a STA-analyzer (Diagnostica Stago). Lipoprotein profiles were determined using ultracentrifugation followed by analysis of the lipoprotein fraction by agarose gel electrophoresis. Apoprotein B levels were measured by nephelometry. Glucose was measured by a glucose hexokinase assay and insulin by competitive radioimmunoassay (Pharmacia).
The insulin-modified frequently sampled IV glucose tolerance test was used to assess insulin sensitivity [13]. After a 10-12 h overnight fast, intravenous cannulas were placed in both antecubital veins. A bolus of a 50% glucose solution (0.3 g/kg) was injected at time 0 and a bolus of regular human insulin (0.03 U/kg) was given after 20 min. Blood samples were collected at -15, -10, -5, -1, 2, 3, 4, 5, 6, 8, 10, 14, 19, 22, 25, 30, 40, 50, 70, 100, 140 and 180 min for determination of plasma glucose and insulin concentrations. The insulin sensitivity index (SI) was calculated from the insulin-modified version of the program MINMOD. The fractional SD (the error in estimating the parameters) for all SI values was < 50%.
Body weight was measured on a calibrated scale with the subjects wearing hospital gowns only. Waist circumference was measured at the level of the umbilicus while the subject was standing and after a normal expiration. Hip circumference was measured at the level of the greatest gluteal protruberence. All waist and hip measurements were performed by the same investigator.
Total fat and lean body mass were determined by dual-energy X-ray absorptiometry (DEXA) using a model DPX-1Q whole body scanner (Lunar Radiation Corp., Madison, Wisconsin, USA). Estimates were made of the amount of fat in the trunk and the extremities. The trunk was defined as that region extending from an upper horizontal border at the lower edge of the chin, lateral borders by vertical lines that bisected each axilla oriented obliquely to include the waist, hip, buttock and thigh tissue to a lower border formed by the intersection of oblique lines extending from the level of the superior aspect of the iliac crest and passing through the hip joint. The arm included the entire shoulder, arm and forearm while the leg included the entire hip, thigh and lower leg.
Computed tomograpy (CT) measurement of abdominal fat: CT was performed on a GE9800 scanner. A scout image was used to approximate a single 10 mm-thick axial image through lumbar vertebrae L4-L5. The image was taken during suspended respiration after a normal expiration. Total abdominal adipose tissue (TAT) area was calculated by delineating the surface with a light pen and then computing the adipose tissue surfaces with an attenuation range of -190 to -30 Hounsfield units. Visceral adipose tissue (VAT) area was measured by drawing a line within the muscle wall surrounding the abdominal cavity. Subcutaneous adipose tissue (SAT) area was calculated by subtracting the VAT area from the TAT area. All CT fat measurements were made by a radiologist who was unaware of the subject's clinical characteristics.
REE was determined by indirect calorimetry using the open circuit technique with the subject in the supine position after a 10-12 h overnight fast. The criterion for a valid metabolic rate was a minimum of 15 min of steady-state, defined as < 10% fluctuation in minute ventilation and oxygen consumption and < 5% fluctuation in respiratory quotient. Metabolic rate was calculated using the Weir equation [14]. The percent of predicted REE was calculated by dividing the measured REE by the metabolic rate predicted from the Harris-Benedict equation [15].
Statistical analyses
One-way analysis of variance was used to compare groups. In most cases, Pearson's product-moment correlation coefficients were used to quantify the associations between adiposity measures and metabolic variables. Spearman rank-order correlations were used when a variable was not normally distributed. Stepwise forward regression analysis was used to measure the contribution of variables to the value of independent variables in a multiple linear regression. All statistical analyses were performed with SigmaStat (Version 2.03, SPSS, Chicago, Illinois, USA) statistical software.
Results
Table 1 shows the characteristics of patients enrolled in the study. The majority of subjects were male. Mean CD4 cell counts and median plasma HIV-1 RNA measurements were similar among the three groups. Body mass index (BMI) was also similar among the groups. Subjects in the PI-LD group tended to be older than those in the two groups without lipodystrophy (P = 0.055). By design, the mean WHR of PI-LD subjects was significantly greater when compared to both PI-controls and non-PI-controls (p < 0.001). Mean waist circumference also was significantly greater in the PI-LD group compared to PI-controls (P = 0.014) and tended to be higher compared to non-PI-controls(P = 0.069). Twelve out of 14 subjects who met criteria for central fat accumulation also had lipoatrophy of the extremities as manifested by venous prominence and a pseudomuscular appearance. The two patients without lipoatrophy of the extremities had facial lipoatrophy manifested by sunken cheeks. In the PI-LD group, seven subjects were taking indinavir, three saquinavir, two nelfinavir and two were taking a ritonavir/saquinavir combination at the time of the study. In the PI-control group, eight subjects were taking indinavir, two nelfinavir, one saquinavir, one ritonavir and one was taking amprenivir. In the PI-LD, PI-controls and non-PI controls, 64%, 54% and 40%, respectively, were receiving stavudine whereas 64%, 77% and 60% were receiving lamivudine, respectively. A small number of subjects in each group were taking zidovudine, zalcitabine, didanosine or abacavir.
Body composition measurements are presented in Table 2. There was no significant difference in DEXA-determined percent body fat, total body fat and lean body mass (LBM) between the three groups. The percent of total body fat located in the trunk as determined by DEXA was significantly greater in the PI-LD group than in the PI-controls and non-PI-controls (P = 0.015 and P = 0.008, respectively). By contrast, the percent of total body fat present in the extremities was significantly lower in the PI-LD group than in the PI-controls (P = 0.037) and the non-PI-controls (P = 0.008). Regional adiposity also differed significantly between groups when measured by CT. VAT was significantly greater in PI-LD subjects than in PI-controls (P = 0.007) and non-PI-controls (P = 0.031), but abdominal SAT area and TAT area were not significantly different between the groups. There were no significant differences in regional adiposity between PI-controls and non-PI-controls.
Metabolic parameters are presented in Table 3. SI was markedly impaired in PI-LD subjects with a mean value of 1.00 × 10-4 (min-1/μU/ml). Mean SI was significantly lower in this group compared with both PI- and non-PI-controls (P < 0.001). Furthermore, PI-controls tended to be less insulin sensitive than non-PI-controls (P = 0.059). Across the entire study population, SI was negatively correlated with measures of central adiposity and positively correlated with percent total body fat present in the extremities (Table 4). There was no significant correlation between SI and total body fat as measured by BMI, percent body fat or total body fat mass or between SI and SAT. In a forward stepwise regression model of SI where the possible predictors were age, BMI, percent body fat, VAT, and percent extremity fat, only VAT was an independent predictor of insulin sensitivity.
Median triglyceride levels tended to be higher in the PI-LD group as compared with the PI-controls (P = 0.053) and non-PI-controls (P = 0.055). Serum triglyceride levels also correlated significantly with central fat accumulation (Table 4). Conversely, triglyceride levels were negatively correlated with percent of total body fat present in the extremities. Mean HDL-C levels did not differ significantly between the groups but serum HDL-C levels were negatively correlated with central adiposity and positively correlated with percent of total body fat present in the extremities (Table 4). In a forward stepwise regression analysis of HDL-C, where the possible predictors were age, SI, triglycerides, percent body fat, BMI, percent extremity fat, and VAT, only VAT was an independent predictor of HDL-C levels. LDL-C and apoprotein B levels did not differ significantly between the groups and did not correlate with measures of total or regional adiposity.
PAI-1 activity and fibrinogen levels also did not differ significantly among the groups. PAI-1 activity tended to be positively correlated with VAT (r, 0.336;P = 0.07). Fibrinogen levels were correlated significantly with percent of total body fat present in the trunk (r, 0.428;P = 0.029) and percent of total body fat present in the extremities (r, -0.525;P = 0.006) but were not correlated with VAT.
REE was significantly higher in PI-LD subjects than in non-PI-controls (P = 0.01;Table 5) and tended to be higher in PI-LD subjects when compared to PI-controls (P = 0.06). The percent predicted REE and REE/kg LBM were significantly higher in the PI-LD group than in the PI- and non-PI-controls, but there was no difference between the control groups. In the PI-LD group, REE was 125% of that predicted by the Harris-Benedict equation versus 106% of predicted REE in PI-controls (P = 0.003) and 96% in non-PI-controls (P < 0.001). Mean REE/kg LBM was significantly greater in PI-LD subjects than in PI-controls (P < 0.001) and non-PI-controls (P < 0.001). Respiratory quotient did not differ significantly between groups. SI was significantly correlated with REE (Table 4) and REE/kg LBM (r, -0.685, P < 0.005). Measures of central adiposity including VAT were also significantly correlated with REE but percent of body fat present in the extremities did not correlate with REE. In a forward stepwise regression analysis, fat free mass (FFM) and SI but not VAT or age were independent predictors of REE.
Discussion
In summary, we found that central adiposity is strongly correlated with the metabolic disturbances associated with the HIV lipodystrophy syndrome and that VAT volume is an independent predictor of insulin sensitivity and HDL-C levels. The percent of total body fat present in the extremities also correlates significantly with metabolic changes, but unlike central abdominal fat, it may protect against metabolic abnormalities. This study also suggests that PIs have an independent metabolic effect on insulin sensitivity. REE is increased in patients with the HIV lipodystrophy syndrome and SI is an independent predictor of REE in this population.
We studied non-obese patients with lipodystrophy manifested by an increase in abdominal girth in addition to loss of fat from at least one depot after the initiation of PI therapy and compared them with PI-treated and PI-naive subjects without any changes in body fat distribution. The majority of patients with central fat accumulation also had evidence of fat loss from the extremities. The presence of visceral adiposity in the PI-LD group was confirmed by CT. PI-treated subjects with central adiposity were significantly less insulin sensitive than PI-treated and PI-naive subjects without changes in body fat distribution. Triglyceride levels tended to be higher in the group with lipodystrophy compared to the two groups without this complication, but HDL-C levels were not significantly different between the groups.
The finding that central adiposity was strongly correlated with metabolic disturbances in the HIV lipodystrophy syndrome was not unexpected. In the general HIV-uninfected population, cross-sectional and longitudinal studies have shown consistently that accumulation of central fat is associated with insulin resistance and a dyslipidemia characterized by elevated triglyceride levels and low HDL-C levels [5-9]. In most studies, VAT is more strongly correlated with metabolic abnormalities than is SAT.
A possible mechanism to explain the association of these metabolic changes with visceral fat accumulation is the high turnover rate of visceral fat [16]. Accelerated release of free fatty acids (FFA) could lead to peripheral and hepatic insulin resistance as well as to an increase in triglyceride and very low density lipoprotein production. FFA did not differ between the groups (data not shown); however, FFA levels may not be a good indicator of FFA turnover [16]. Support for a causal role of visceral fat in insulin resistance comes from an obese rat model in which surgical removal of intraabdominal fat depots reverses hepatic insulin resistance and decreases serum insulin levels [17]. According to an alternative model, insulin resistance with hyperinsulinemia could lead to visceral obesity [18]. In this model, muscle-specific insulin resistance leads to hyperinsulinemia which increases insulin action in adipose tissue. Because insulin increases lipid storage and decreases lipolysis, fat accumulation is favored. This effect is especially pronounced in visceral adipose tissue as opposed to subcutaneous depots because of its higher cellularity and blood flow.
Previous studies of HIV lipodystrophy have shown an association between body fat changes and metabolic disturbances [1,10,19]. In the first study to describe the lipodystrophy syndrome, higher triglyceride, insulin and C-peptide levels and more severe insulin resistance were found in PI-treated patients with lipodystrophy than in PI-treated patients without lipodystrophy [1]. In the LIPOCO study [20], VAT area measured by CT was positively correlated with fasting insulin levels. However, other studies suggest that metabolic changes and fat redistribution are unrelated. A study that compared metabolic parameters in HIV-infected subjects before and after approximately 3 months of PI-based therapy found that triglyceride, LDL-C, glucose and insulin levels increased without any changes in body fat distribution as measured by DEXA [21]. Another study examined the effects of ritonavir on lipid levels in 21 HIV-uninfected volunteers [22]. Administration of ritonaivr for 2 weeks significantly increased serum triglyceride and total cholesterol levels and lowered HDL-C levels. Body weight was unchanged over this period of time. A third study measured changes in carbohydrate metabolism, lipid profiles and body composition in 10 HIV-uninfected men after 4 weeks of treatment with indinavir at a dose of 800 mg three times daily [11]. No significant changes in total cholesterol, LDL-C and HDL-C or triglyceride levels were noted, but insulin-mediated glucose disposal rate (measured by euglycemic clamp) decreased significantly in the absence of body fat changes as measured by CT and DEXA.
Although the present study suggests that visceral fat accumulation is an important determinant of the metabolic abnormalities observed in the HIV lipodystrophy syndrome, the loss of peripheral fat may also play a role. The strong correlation between percent of total body fat present in the extremities and metabolic disturbances observed in this study suggests that peripheral fat may play a protective role in the HIV lipodystrophy syndrome, even though extremity fat was not an independent predictor of metabolic changes. WHR had the strongest correlation with SI in this study, perhaps because WHR reflects both central fat accumulation and peripheral fat loss in patients with HIV-associated lipodystrophy. In a recent study of HIV-infected patients with lipodystrophy, midthigh circumference was a significant predictor of fasting insulin levels, but patients with lipoatrophy alone were more insulin sensitive than those with both central fat accumulation and peripheral wasting [10]. In another study, loss of limb fat but not accumulation of truncal fat was strongly associated with insulin resistance [23]. The peripheral fat wasting that often accompanies central fat accumulation may therefore have an additive effect on the metabolic disturbances associated with the HIV lipodystrophy syndrome. The metabolic importance of subcutaneous fat is also suggested by the lipodystrophies encountered in the general population, where atrophy of subcutaneous fat is associated with insulin resistance, diabetes mellitus and elevated triglyceride levels [24,25]. Overall, the results of this study suggest that PI-associated changes in metabolism may in part be dependent on changes in body fat distribution. However, these drugs also appear to have independent metabolic effects.
The third important finding of our study was the novel observation that REE, REE/kg LBM and percent predicted REE were significantly greater in the group with lipodystrophy than in those without lipodystrophy and that insulin sensitivity and REE were strongly and negatively correlated. Furthermore, in a forward stepwise regression model, SI was an independent determinant of REE in addition to LBM. Increased REE has been noted in several but not all studies of HIV-infected men and women with and without weight loss in both the pre-highly active antiretroviral therapy (HAART) [26-28] and HAART eras [29,30], but the mechanism of this increase is not known. In the pre-HAART era, increased REE may have been secondary to high levels of inflammatory cytokines associated with uncontrolled HIV infection. In the HAART era, increased REE probably has a different etiology. An earlier study showed that HAART administration is independently associated with increased REE in HIV-infected patients [30]. The present study confirms that increased REE in HIV infection persists in the HAART era, and suggests that it may be a feature of the HIV lipodystrophy syndrome even in patients with good virologic control.
Increased REE has also been found in otherwise healthy subjects with impaired glucose tolerance and diabetes mellitus. For example, the sleeping metabolic rate in Pima Indians with impaired glucose tolerance and diabetes mellitus is 2.7 and 4.9% higher, respectively, than controls with normal glucose tolerance [31]. Endogenous glucose output, fasting insulin and FFA concentrations and insulin-stimulated glucose disposal were each significant determinants of REE independent of established determinants such as FFM. Mechanisms proposed to explain these findings include various energy-consuming substrate cycles such as gluconeogenesis and insulin-induced increases in sympathetic nervous system activity.
Elevated basal metabolic rates have also been described in patients with various acquired and congenital forms of lipodystrophy [24,32-34]. The mechanism of increased REE in these lipodystrophies is not known, but in many cases, affected patients are insulin resistant. As the major part (80%) of whole-body REE stems from the metabolic activity of the liver, kidneys, heart, brain and skeletal muscle [35], it is likely that the high REE encountered in HIV lipodystrophy and other lipodystrophies is due to hypermetabolism in one or more of these tissues.
The results of this study have several implications. First, the strong associations between body fat distribution and insulin resistance in the HIV lipodystrophy syndrome suggest causality - but in which direction? Central adiposity in this syndrome may lead to insulin resistance via accelerated release of FFA. On the other hand, insulin resistance with hyperinsulinemia may lead to visceral fat accumulation. The rapid development of insulin resistance on PI therapy described in the studies mentioned above suggests that the development of insulin resistance precedes visceral lipohypertrophy in the HIV lipodystrophy syndrome and may therefore contribute to accumulation of lipid in this depot. Prospective studies in patients initiating PI-based therapy are needed to understand better the link between insulin resistance and visceral fat accumulation. The strong associations between fat redistribution and metabolic changes also suggest that the metabolic disturbances associated with PI therapy may not reverse completely if PI are discontinued but body composition remains altered. Given the strong correlations between both central fat accumulation and peripheral lipoatrophy and metabolic disturbances, patients with peripheral lipoatrophy alone and with central obesity alone need to be included in future studies in order to determine the individual influences of fat atrophy and fat hypertrophy on metabolic parameters.
In conclusion, both visceral lipohypertrophy and peripheral lipoatrophy are strongly associated with metabolic disturbances in the HIV lipodystrophy syndrome. The accumulation of central abdominal fat and the loss of peripheral subcutaneous fat may each have important metabolic consequences in patients on antiretroviral therapy. However, there is evidence that PI also have independent effects on metabolic parameters. The finding that PI-naive subjects tend to be more insulin sensitive than PI-treated subjects without lipodystrophy lends further support to this idea. Finally, resting energy expenditure is increased in HIV lipodystrophy and is highly correlated with insulin sensitivity.
Acknowledgements
The authors thank the patients who participated in this study, the nurses and the personnel of the CORE laboratory and nutrition service of the General Clinical Research Center, B. Young and K. Greenberg, and D. Bessesen for review of the manuscript.
References
1. Carr A, Samaras K, Burton S. et al. A syndrome of peripheral lipodystrophy, hyperlipidemia and insulin resistance due to HIV protease inhibitors. AIDS 1998, 12: F51-F58.
2. Saint-Marc T, Partisani M, Poizot-Martin I. et al. A syndrome of peripheral fat wasting (lipodystrophy) in patients receiving long-term nucleoside analoque therapy. AIDS 1999, 13: 1659-1667.
3. Gervasoni C, Ridolfo AL, Trifiro G. et al. Redistribution of body fat in HIV-infected women undergoing combined antiretroviral therapy. AIDS 1999, 13: 465-471.
4. Safrin S, Grunfeld C. Fat distribution and metabolic changes in patients with HIV infection. AIDS 1999, 13: 2494-2505.
5. Marcus MA, Murphy L, Pi-Sunyer FX, Albu JB. Insulin sensitivity and serum triglyceride level in obese white and black women: relationship to visceral and truncal subcutaneous fat. Metabolism 1999, 48: 194-199.
6. Ross R, Fortier L, Hudson R. Separate associations between visceral and subcutaneous adipose tissue distribution, insulin and glucose levels in obese women. Diabetes Care 1996, 19: 1404-1411.
7. Pouliot M-C, Despres J-P, Nadeau A. et al. Visceral obesity in men: associations with glucose tolerance, plasma insulin, and lipoprotein levels. Diabetes 1992, 41: 826-834.
8. Ronnemaa T, Koskenvuo M, Marniemi J. et al. Glucose metaboism in indentical twins discordant for obesity: the critical role of visceral fat. J Clin Endoc Metab 1997, 82: 383-387.
9. Lemieux S, Prud'homme D, Nadeau A, Tremblay A, Bouchard C, Despres JP. Seven-year changes in body fat and visceral adipose tissue in women. Diabetes Care 1996, 19: 983-991.
10. Hadigan C, Meigs JB, Corcoran C. et al. Metabolic abnormalities and cardiovascular disease risk factors in adults with human immunodeficiency virus infection and lipodystrophy. Clin Infect Dis 2001, 32: 130-139.
11. Noor M, Lo J, Schwarz JM. et al. Metabolic effects of indinavir in healthy HIV-seronegative men. AIDS 2001, 15: F11-F18.
12. Hatano H, Miller KD, Yoder CP. et al. Metabolic and anthropometric consequences of interruption of highly active antiretroviral therapy. AIDS 2000, 14: 1935-1942.
13. Saad MF, Anderson RL, Laws A. et al. A comparison between the minimal model and the glucose clamp in the assessment of insulin sensitivity across the spectrum of glucose tolerance. Diabetes 1994, 43: 1114-1121.
14. Weir JBV. New methods for calculating metabolic rate with special reference to protein metabolism. J Appl Physiol 1953, 6: 317-334.
15. Harris JA, Benedict FG. A Biometric Study of Basal Metabolism in Man. Washington, DC: Carnegie Institute of Washington, 1919.
16. Coppack SW, Jensen MD, Miles JM. In vivo regulation of lipolysis in humans. J Lipid Res 1994, 35: 177-193.
17. Barzilai N, Liu B-Q, Vugiun P, Cohen P, Wang J, Rossetti L. Surgical removal of visceral fat reverses hepatic insulin resistance. Diabetes 1999, 48: 94-98.
18. Bjorntorp P. The regulation of adipose tissue distribution in humans. Int J Obes 1996, 20: 291-302.
19. Carr A, Samaras K, Thorisdottir A, Kaufmann GR, Chisholm DJ, Cooper DA. Diagnosis, prediction, and natural course of HIV-1 protease inhibitor-associated lipodystrophy, hyperlipidemia and diabetes mellitus: a cohort study. Lancet 1999, 353: 2093-2099.
20. Saint-Marc T, Partisani M, Poizot-Martin I. et al. Fat distribution evaluated by computed tomography and metabolic abnormalities in patients undergoing antiretroviral therapy: preliminary results of the LIPOCO study. AIDS 2000, 14: 37-49.
21. Mulligan K, Grunfeld C, Tai VW. et al. Hyperlipidemia and insulin resistance are induced by protease inhibitors independent of changes in body composition in patients with HIV infection. J Acquir Immune Def Syndr 2000, 23: 35-43.
22. Purnell JQ, Zambon A, Knopp RH. et al. Effect of ritonavir on lipids and post-heparin lipase activities in normal subjects. AIDS 2000, 14: 51-57.
23. Mynarcik DC, McNurlan MA, Steigbigel RT, Fuhrer J, Gelato MC. Association of severe insulin resistance with both loss of limb fat and elevated tumor necrosis factor receptor levels in HIV lipodystrophy. J Acquir Immune Def Syndr, 25:312-321.
24. Klein S, Farook J, Wolfe RR, Stuart CA. Generalized lipodystrophy: in vivo evidence for hypermetabolism and insulin-resistant lipid, glucose and amino acid kinetics. Metabolism 1992, 41: 893-896.
25. Robbins DC, Horton ES, Tulp O, Sims EAH. Familial partial lipodystrophy: complications of obesity in the non-obese? Metabolism 1982, 31: 445-452.
26. Grunfeld C, Pang M, Shimizu L, Shigenaga JK, Jensen P, Feingold KR. Resting energy expenditure, caloric intake, and short-term weight change in human immunodeficiency virus infection and the acquired immunodeficiency syndrome. Am J Clin Nutr 1992, 55: 455-460.
27. Hommes MJT, Romijn JA, Godfried MH. et al. Increased resting energy expenditure in human immunodeficiency virus-infected men. Metabolism 1990, 39: 1186-1190.
28. Suttmann U, Ockenga J, Hoogestraat L. et al. Resting energy expenditure and weight loss in human immunodeficiency virus-infected patients. Metabolism 1993, 42: 1173-1179.
29. Grinspoon S, Corcoran C, Miller K. et al. Determinants of increased energy expenditure in HIV-infected women. Am J Clin Nutr 1998, 68: 720-725.
30. Shevitz AH, Knox TA, Spiegelman D, Roubenoff R, Gorbach SL, Skolnik PR. Elevated resting energy expenditure among HIV-seropositive persons receiving highly active antiretroviral therapy. AIDS 1999, 13: 1351-1357.
31. Weyer C, Bogardus C, Pratley RE. Metabolic factors contributing to increased resting metabolic rate and decreased insulin-induced thermogenesis during the development of type 2 diabetes. Diabetes 1999, 48: 1607-1614.
32. Cutler DL, Kaufmann S, Freidenberg GR. Insulin-resistance diabetes mellitus and hypermetabolism in mandibuloacral dysplasia: a newly recognized form of partial lipodystrophy. J Clin Endocrinol Metab 1991, 73: 1056-1061.
33. Kobberling J, Willlms B, Kattermann R, Creutzfeldt W. Lipodystrophy of the extremities: A dominantly inherited syndrome associated with lipoatrophic diabetes. Humangenetick 1975, 29: 111-120.
34. Jensen MD. Adrenergic regulation of lipolysis in a patient with lipoatrophy of the upper body. Mayo Clin Proc 1991, 66: 704-710.
35. Elia M. Organ and tissue contribution to metabolic rate. In Energy Metabolism: Tissue Determinants and Cellular Corrollaries. Edited by Kinney JM, Tucker HN. New York: Raven Press; 1992: 61-79.
© 2001 Lippincott Williams & Wilkins, Inc.