Improved Insulin Sensitivity and Body Fat Distribution in HIV-Infected Patients Treated With Rosiglitazone: A Pilot Study
Summary: The insulin-sensitizing drugs thiazolidinediones (TZDs), such as rosiglitazone, improve insulin sensitivity and also promote adipocyte differentiation in vitro. The authors hypothesized that TZDs might be beneficial to patients with HIV disease to improve insulin sensitivity and the distribution of body fat by increasing peripheral fat. The ability of rosiglitazone (8 mg/d) to improve insulin sensitivity (from hyperinsulinemic-euglycemic clamp) and to improve body fat distribution (determined from computed tomography measurements of visceral adipose tissue [VAT] and subcutaneous adipose tissue [SAT]) was determined in 8 HIV-positive patients. Before treatment, the insulin sensitivity of the patients was reduced to approximately 34% of that in control subjects. The rate of glucose disposal during a hyperinsulinemic-euglycemic clamp (Rd) was 3.8 ± .4 (SEM) mg glucose/kg lean body mass/min compared with 11.08 ± 1.1 (p < .001) in healthy age- and body mass index (BMI)-matched control subjects. After rosiglitazone treatment of 6 to 12 weeks, Rd increased to 5.99 ± .9 (p = .02), an improvement of 59 ± 22%. SAT increased by 23 ± 10% (p = .05), and, surprisingly, VAT was decreased by 21 ± 8% (p = .04) with a trend for increased SAT/VAT that failed to reach statistical significance. There were no significant changes in blood counts, viral loads, or CD4 counts with rosiglitazone treatment. The study demonstrates that rosiglitazone therapy improves insulin resistance and body fat distribution in some patients with HIV disease.
JAIDS Journal of Acquired Immune Deficiency Syndromes 2002; 31(2):163-170 (October 02)
Marie C. Gelato; Departments of Medicine, Radiology, and Surgery, State University of New York at Stony Brook, Stony Brook, New York, U.S.A.
Thiazolidinedione therapy should, in theory, be a good option to improve insulin sensitivity and body fat distribution. In the IR associated with DM, the TZDs restore the ability of insulin to suppress hepatic glucose output and increase peripheral glucose disposal (18). In addition, these agents stimulate the differentiation of adipocytes from the peripheral compartment more than those from the visceral compartment, which may lead to an increase in peripheral body fat (19). Although Maggs et al. (20) reported improved insulin sensitivity in subjects with type 2 diabetes treated with troglitazone and Walli et al. (16) have reported improved insulin sensitivity in an HIV-positive population with type 2 DM, troglitazone (Rezulin) is no longer available due to serious liver toxicity associated with its use (21). Rosiglitazone is a newer thiazolidinedione with less liver toxicity and fewer drug-drug interactions (22). Therefore, rosiglitazone may be a better therapeutic agent for HIV-positive patients.
This pilot study reports on the first 8 HIV-positive patients with IR treated with rosiglitazone (8 mg daily) for 6 to 12 weeks. Insulin sensitivity and body fat distribution were assessed before and after therapy.
Nine HIV-positive patients with IR documented by a hyperinsulinemic-euglycemic clamp (23) were recruited into the study. Subjects were classified as IR if their measured insulin sensitivity was at least 1 SD below the mean value for age- and body mass index (BMI)-matched control subjects.
Patients received 4 mg of rosiglitazone (Avandia, GlaxoSmithKline, Reseach Triangle Park, NC, U.S.A.) twice daily for 6 to 12 weeks. They were monitored every 2 weeks for liver function tests, complete blood counts, CD4 counts, and viral load. In addition, blood pressure, weight, and physical appearance were noted at each
visit. One patient was dropped from the study because of persistent abnormalities of liver function that returned to normal after discontinuing rosiglitazone. Thus, 8 patients completed at least 6 weeks of therapy (1 patient, 6 weeks; 7 patients, 12 weeks). The patient who stopped therapy at 6 weeks did so because of unacceptable weight gain. At the end of the treatment period, all patients were again assessed for insulin sensitivity, lean body mass, VAT and SAT, fasting lipid profiles, waist-to-hip ratio, blood pressure, viral load, CD4 count, complete blood count, liver function tests, and serum chemistries. Although changes in body fat distribution were assessed before and after therapy with rosiglitazone, altered body fat distribution was not a specific criterion for inclusion. Antiretroviral therapies remained unchanged throughout the study period. Although subjects were asked not to alter their exercise regimens during the time they were taking rosiglitazone, subjects were not confined to a metabolic facility for the entire 6 to 12 weeks.
Insulin sensitivity was also measured in 10 healthy seronegative control subjects. The control subjects were closely matched to the HIV-infected persons for age, BMI, gender, and ethnicity. The control subjects included 5 men and 5 women, aged 40 ± 2 years, with a BMI of 26 ± 1. The control subjects were 70% white, 20% black, and 10%
Hispanic. None of the control subjects had first-degree relatives with DM. None was taking chronic medications, and none had abnormal thyroid or gonadal function. The HIV-positive subjects included 3 men and 5 women, aged 41 ± 3 years, with a BMI of 26 ± 2. Ethnically, the group of HIV-infected subjects was 75% white, 12.5% black, and 12.5% Hispanic.
Insulin sensitivity was determined as the rate of infused glucose necessary to maintain euglycemia during an infusion of insulin (hyperinsulinemic-euglycemic clamp) (23). Patients were admitted to the General Clinical Research Center the
night before the study, fed a uniform snack at 9:00 p.m., and then fasted from midnight. At 7:00 a.m. after baseline sampling, the subjects were infused with 1.2 mU insulin (Humulin, Eli Lilly)/kg body weight per minute to elevate plasma insulin levels to [sim]40 [mu]U/mL, sufficient to suppress hepatic glucose production in other insulin-resistant states (24). Dextrose was administered intravenously at variable rates to maintain plasma glucose at 90 mg/dL. Plasma glucose was assessed from arterialized blood, obtained by the heated hand technique (25). IR was determined between
the second and third hour of insulin infusion. To normalize for differences in body composition, insulin sensitivity is expressed as mg glucose/kg lean body mass (LBM) per minute. Subjects were classified as IR if their rate of glucose disposal was at least 1 SD below the mean value for age- and BMI-matched control subjects.
Lean body mass was determined by dual-energy x-ray absorptiometry (DEXA) performed with a total body scanner (model DPS Lunar Radiation Co., Madison, WI, U.S.A.). The percentage of total body fat that was present in the limbs was calculated from total fat kg and total limb fat from DEXA.
A computed tomography (CT) scan with a single helical scanner (HiSpeed CT/i, GE Medical Systems, Milwaukee, WI, U.S.A.) without contrast has been used to estimate VAT and SAT at the level of the umbilicus or the fourth lumbar vertebra (4, 6, 26). A trackball-controlled planimetry cursor was used to measure the tissue compartments. The area of the adipose tissue in each compartment was determined by manufacturer-supplied software, which sums the area of pixels in the digital image with CT values between -50 and -150 Hounsfield units that corresponds to adipose tissue. SAT (cm2) and total adipose tissue were measured directly, and the VAT (cm2) was calculated by difference between total adipose tissue and SAT.
The characteristics for the HIV-positive patients are listed in Table 1. All patients had well-controlled disease with either low or undetectable viral loads, and CD4 counts were greater than 300 in 7 of the 8 subjects. Five of the 8 patients were taking the PI indinavir from 6 months to 5 years. The three patients not on PI therapy at the time of study had previously been on PIs for 3 to 5 years, but the PIs had been discontinued 6 months before entering the study. Two patients (patients 2 and 6) were taking lipid-lowering agents (atorvastatin, gemfibrozil). The fasting glucose levels for HIV-positive subjects were 108 ± 5 mg/dL, not significantly different from control subjects (99 ± 3) of comparable age (40 ± 2 years) and BMI (26 ± 1 kg/m2). One HIV-positive patient (patient 1) admitted into the study with a screening glucose of 116 had a fasting glucose above 126 mg/dL. Free fatty acid levels were .43 ± .06 mmol/L, which was similar to the mean level for control subjects (.37 ± .03). Triglyceride levels were significantly higher in HIV-positive subjects (366 mg/dL ± 65) than in control subjects (93 ± 15, p < .001). HDL levels were significantly
lower in the HIV-positive patients (34 ± 2.5 mg/dL) than in control subjects (56 ± 4, p < .001). In the 6 HIV-positive patients in whom triglyceride levels did not preclude the assessment of LDL, LDL levels were 97 ± 18 compared with 109 ± 8 (p = NS) in control subjects.
There was no difference in the amount of body fat assessed from DEXA in control subjects and HIV-LD subjects when fat mass was normalized for height. The total fat in the control group was 7.1 ± 1.0 kg/m2 and 6.0 ± 1.1 (NS, p = .5) in the HIV-LD group. There was, however, a difference in body fat distribution between the HIV-positive patients and the control subjects. The amount of body fat in the limbs relative to total body fat was significantly different between control subjects (48.3% ± 1.1) and those with HIV-LD (30.9% ± 2.33, p < .001).
Estimates of body fat were also determined from the cross-sectional area of a CT scan of the abdomen at the fourth lumbar vertebra because there is improved accuracy from CT scanning (e.g., 6,27,28). Both SAT and VAT were estimated in HIV-infected subjects and in healthy control subjects. The mean ratio of SAT to VAT was 1.2 ± .23 in HIV-positive patients and 2.73 ± .35 in control subjects, suggesting less subcutaneous fat and more visceral fat in patients with HIV disease (p = .02). The difference in body fat distribution between control and HIV-positive subjects
was also apparent in the less precise measure of waist-to-hip ratio. The mean value of the waist-to-hip ratio for HIV-positive patients was .95 ± .02, while that of control subjects was significantly less, .81 ± .02 (p < .001).
At baseline, all 8 patients had values for insulin sensitivity that were only 34% of the values in the control subjects. Insulin sensitivity was assessed with a hyperinsulinemic-euglycemic clamp study with infusion of insulin to achieve
plasma concentrations in the normal, physiologic range ([sim]40 [mu]U/mL). Insulin-stimulated glucose disposal rates (Rd) for the HIV group were 3.8 ± 0.4 mg glucose/kg LBM/min compared with 11.08 ± 1.1 for the control group (p < .001). After 12 weeks of rosiglitazone therapy, insulin sensitivity increased significantly to 5.99 ± .9 mg/kg LBM/min (p = .02), an increase of 59 ± 22%. After taking rosiglitazone, the insulin sensitivity of one patient was unchanged, and one patient did not have a clamp study after rosiglitazone therapy.
After rosiglitazone therapy, SAT increased, with a mean increase of 23 ± 10% (130 ± 54 cm2, p = .046. Interestingly, VAT decreased by 21 ± 8% (-148 ± 59 cm2, p = .04). Although the SAT-to-VAT ratio increased in all patients after rosiglitazone therapy, the variability of the response prevented the difference (.74 ± .4) from reaching statistical significance (p = .1. Three subjects reported a perceptible change in body fat distribution. There was also a tendency for LBM to increase, with the mean difference in all patients after rosiglitazone treatment of 1.6 ± 0.73 kg (p = .06).
Improved sensitivity to insulin with increased SAT and decreased VAT after rosiglitazone therapy was not accompanied by significant changes in fasting glucose, triglycerides, or free fatty acids (Table 2), but fasting insulin levels declined
significantly (p = .03). Although there was no consistent change in the circulating levels of triglycerides, 3 subjects did show an increase in triglyceride levels with rosiglitazone. One patient was withdrawn from rosiglitazone due to an elevation in liver function tests. For the remaining 8 subjects, there were no demonstrable safety
concerns with rosiglitazone therapy. There were no significant changes in hemoglobin and hematocrit, serum chemistries, viral load, or CD4 cell counts during rosiglitazone administration. One patient with a history of decreased platelet levels had a decreased platelet count with rosiglitazone therapy. With discontinuation of rosiglitazone, the platelet count returned to baseline within 2 to 3 days. Four of the 8 patients gained 1 to 3 kg, but this was not manifest as peripheral edema in any of the patients.
We have previously documented IR in patients with HIV disease who did not have overt DM (5). We (5), and others (29), have demonstrated that IR was associated with the loss of SAT and preservation of, or increased, VAT. This current study demonstrates, for the first time, that rosiglitazone therapy substantially improves insulin sensitivity (mean increase, 59 ± 22%, Fig. 2 and Table 2) in patients with HIV disease. This demonstration of improved insulin sensitivity was observed with a hyperinsulinemic-euglycemic clamp technique (23) at physiologic levels of insulin. In
contrast, the clinical trials demonstrating improvement in insulin sensitivity in patients with type 2 DM with troglitazone therapy were carried out with about threefold higher levels of infused insulin (120 mU/m2/min [(20)] vs. [sim]40
mU/m2/min in the current study).
Although the current study demonstrates a highly significant improvement in insulin sensitivity overall, there was one subject (subject 7) who had no improvement in insulin sensitivity. Non-responders to troglitazone therapy have also been reported by Walli et al. (16), in HIV-positive patients, and by Suter et al. (30), in diabetic patients. In addition, the TZDs troglitazone (20) and rosiglitazone (31, 32) used as monotherapy normalize fasting glucose and glycosylated hemoglobin in only about 20% to 30% of patients with type 2 DM. So although rosiglitazone improved insulin
sensitivity in HIV-infected persons with lipodystrophy, insulin sensitivity did not return to control levels after 6 to 12 weeks of treatment. This is in keeping with other studies of TZD therapy, which indicate an improvement in glucose
dynamics but not complete normalization in all patients.
Metformin therapy has also been demonstrated to improve insulin sensitivity in HIV-positive subjects with IR (15, 33). Metformin decreased insulin levels by 20% as assessed with an oral glucose tolerance test but did not significantly alter
glucose levels (15). With metformin therapy, there was loss of VAT, but, unfortunately, this was also accompanied with a loss of SAT (15, 33). Overall, loss of body fat may be advantageous for patients with type 2 DM, but in those
with HIV disease, where loss of SAT is already a problem, further loss of SAT is not desirable.
This study tested the hypothesis that TZDs, such as rosiglitazone, might be ideally suited as a therapeutic strategy in HIV-positive patients to improve not only insulin sensitivity but also improve distribution of body fat. Although the exact mechanism by which TZDs improve insulin sensitivity is not known, there is evidence from animal studies that they can enhance insulin sensitivity by decreasing the effects of tumor necrosis factor-[alpha] (TNF[alpha]) (34). Further, we have shown that insulin sensitivity was inversely related to the type 2 soluble receptor for TNF[alpha] (5).
TZDs are ligands for the gamma form of the peroxisome proliferator-activated receptor (PPAR[gamma]), a nuclear hormone receptor highly expressed in adipose tissue (35). Moreover, the effects of TZDs in humans may be depot-specific, as PPAR expression is greater in SAT than in VAT (19). Walli et al. have investigated the ability of the TZD troglitazone to improve insulin sensitivity and body fat redistribution in HIV-positive patients with type 2 DM (16). Although there was a demonstrable improvement in insulin sensitivity, the effect of troglitazone on body fat distribution
was more equivocal. Two subjects lost SAT and VAT; two subjects gained SAT and VAT; and two subjects gained SAT but lost VAT (16). The ability of troglitazone therapy to decrease VAT and increase SAT has been shown in subjects with type 2 DM (36-39). However, rosiglitazone may be a better choice of therapy since troglitazone therapy has been associated with liver toxicity and withdrawn for patients with type 2 DM (21). Thus, this study examined the ability of rosiglitazone to improve insulin sensitivity and body fat distribution in HIV-infected persons with reduced insulin
Subjects were recruited into this study on the basis of reduced insulin sensitivity. While it was not an inclusion criterion for the study, the HIV-infected subjects also had an altered distribution of body fat with significantly less of their body fat in the limbs assessed with DEXA compared with control subjects (31% ± 2 vs. 48% ± 1, p < .001). However, DEXA scanning alone did not allow sufficient sensitivity to quantitate changes in body fat with rosiglitazone therapy. We, and others (6), have found SAT and VAT determinations from CT provide greater sensitivity than DEXA for assessing body fat distribution. Many of these HIV-positive persons with IR in this present study also had decreased SAT and increased VAT (Table 3). An improvement in insulin sensitivity with rosiglitazone therapy was accompanied by an improvement in fat distribution, with decreased VAT and increased SAT (Table 3). An increase in SAT and decreased VAT has also been reported in troglitazone-treated subjects with non-HIV-associated lipodystrophy (40), although a recent preliminary report of rosiglitazone treatment of subjects with HIV-LD did not confirm changes in either SAT or VAT assessed with magnetic resonance imaging (MRI) (41).
The precise mechanism for the improvement in body fat distribution from rosiglitazone therapy suggested by the current study was not investigated, but there are two possibilities. Improvement in insulin sensitivity alone may be sufficient to alter visceral and truncal fat. This would be in keeping with data indicating loss of VAT with metformin therapy in patients with HIV (15). However, improved insulin sensitivity does not explain the increase in SAT observed in the current study with rosiglitazone treatment. Carr et al. (29) had postulated that PIs might cause loss of peripheral fat by interfering with the PPAR[gamma] receptor system to alter adipocyte differentiation, leading to adipose redistribution. Since the TZDs are ligands for PPAR[gamma], they may be able to reverse some of the effect of PIs. However, since patients on therapies without PIs are also developing this fat redistribution syndrome, clearly more than one mechanism may be involved. Importantly, a recent study by Carr et al. demonstrated that the elimination of PIs from the therapeutic regimen failed to increase peripheral fat accumulation or to improve insulin sensitivity (42).
While rosiglitazone therapy may provide significant benefit, there are some disadvantages to this type of therapy. Patients need to be monitored for liver function abnormalities. Rosiglitazone did cause weight gain in half the patients.
The weight gain may be due, in part, to increased SAT. While rosiglitazone can cause edema, this was not a problem in the current study. Since LBM was measured with the DEXA, it is not possible to distinguish increased LBM from increased total body water. Therefore, rosiglitazone should not be given to patients who have significant heart disease, especially a history of congestive heart failure. So while it has some distinct advantages over metformin and GH therapies, it is not an ideal therapy. It would certainly be more advantageous to have a TZD that had the same effects as
rosiglitazone on insulin sensitivity and body composition with a more potent effect on lipids and no weight gain. Pioglitazone, another TZD, does have more potent lipid-lowering effects; however, it also causes significant weight gain
In conclusion, rosiglitazone therapy performed as predicted. It significantly improved insulin sensitivity, decreased VAT, and increased SAT. It was well tolerated and in this small cohort had no major side effects. Thus, rosiglitazone deserves more extensive study in this population of patients, for which it has potential as a therapeutic agent for the current metabolic complications of HAART.