Secondary Logo

Journal Logo

CLINICAL SCIENCE

Cumulative exposure to nucleoside analogue reverse transcriptase inhibitors is associated with insulin resistance markers in the Multicenter AIDS Cohort Study

Brown, Todd Ta; Li, Xiuhongb; Cole, Stephen Rb; Kingsley, Lawrence Ac; Palella, Frank Jd; Riddler, Sharon Ac; Chmiel, Joan Sd; Visscher, Barbara Rc,e; Margolick, Joseph Bb; Dobs, Adrian Sa

Author Information
doi: 10.1097/01.aids.0000181011.62385.91
  • Free

Abstract

Introduction

Among HIV-infected persons receiving antiretroviral therapy (ART), diminished insulin sensitivity is common and has been associated with dyslipidemia, central adiposity, and peripheral fat wasting [1,2]. Although the pathogenesis of insulin resistance in HIV-infected patients has not been defined, direct and indirect effects of antiretroviral medications likely play a role. Abnormalities in adipose tissue metabolism in HIV-infected persons receiving ART include increased intramyocellular lipid accumulation [3], changes in adipocytokine secretion [4–6], and altered adipose gene expression [7]; all of which have also been associated with reduced peripheral insulin sensitivity. Direct effects of certain antiretroviral medications on insulin sensitivity has been convincingly demonstrated in studies of healthy volunteers [8,9]. The protease inhibitor (PI) indinavir reduced insulin sensitivity by 34% after a single dose [8], and lopinavir/ritonavir administration for 5 days was associated with a 24% reduction in insulin sensitivity [9].

Because abnormalities in glucose metabolism among HIV-infected patients were recognized soon after the introduction of PIs in 1996 [10], this class of medications has been actively investigated. Multiple cross-sectional studies have shown reduced insulin sensitivity in HIV-infected patients exposed to PIs compared with HIV-infected patients receiving ART without a PI [3,11]. However, PIs appear to be heterogeneous in their effects on glucose metabolism [12] and fasting hyperinsulinemia, a marker of insulin resistance, has been reported in HIV-infected patients both treated and untreated with a PI [13,14], suggesting that other medications or factors related to HIV infection itself may be important. In a case–control study of 71 HIV-infected patients with lipodystrophy and 90 HIV-negative controls, Hadigan and colleagues [1] found that duration of exposure to nucleoside reverse transcriptase inhibitors (NRTI), but not PIs, was independently associated with fasting hyperinsulinemia.

The present study used fasting surrogate measures to compare insulin resistance in a large cohort of HIV-infected men having heterogeneous ART exposures with a similar HIV-seronegative group. In addition, the study sought to identify factors, both related and unrelated to ART, which may be associated with markers of insulin resistance. Previous studies examining these relations have been relatively small [15], have not used an internal HIV-seronegative comparison group [1], and have not used extensive, prospectively collected ART data [1,15,16].

Methods

Study participants

The Multicenter AIDS Cohort Study (MACS) is an ongoing prospective study that enrolled 5622 homosexual and bisexual men between 1984 and 1991 at sites located in Pittsburgh, Baltimore, Chicago, and Los Angeles. Institutional review boards at each site approved the protocol and forms, and each participant gave written informed consent. Details of the study design and methods have been published [17].

Briefly, participants attend semiannual study visits that include a detailed interview, physical examination, and collection of biological specimens. Beginning with the 31st semiannual study visit, between April and October 1999, a fasting (≥ 8 h) serum sample was obtained.

Of the 5622 men enrolled in MACS, 1857 of 2885 HIV-seronegative men were administratively censored in 1996 and 1750 men had died by 1 April 1999, leaving 2015 men. Of these, 1799 (89%) had a study visit between 1 April 1999 and 31 March 2003; 1408 of them had a fasting serum sample from which glucose and insulin concentrations were determined. The first visit at which a participant underwent a fasting insulin determination was defined as the index visit. A group of 120 men were excluded because they reported at the index visit use (within 6 months) of medications that affect insulin levels, including antidiabetic medications (i.e., insulin, sulfonylureas, thiazolidinediones, biguanides, meglitinides, or alpha-glucosidase inhibitors) or glucocorticoids. With this exclusion, the final study sample comprised 1288 men. Men who used these medications during follow-up were censored at the last visit prior to starting these medications.

Endpoint ascertainment

Two endpoints were examined at each visit, based on fasting glucose and insulin concentrations. The first was QUICKI, defined as 1/[log10(insulin) + log10(glucose)] [18], and multiplied by 100 for easier interpretation. A lower level suggests more insulin resistance. The second endpoint was a binary outcome of whether or not fasting insulin was > 15 μU/ml at each visit. Fasting insulin levels were measured using a radioimmunoassay technique [coefficient of variation (CV), 2.6%; Linco Research, St Charles, Missouri, USA]. This cut-off point was based on the published normal range for the insulin assay used. Fasting glucose levels were measured by the combined hexokinase/glucose-6-phosphate dehydrogenase method [19] (CV, 1.8%). Both assays used serum samples that had been stored at −80°C and were processed at a central laboratory (Heinz Laboratory, Pittsburgh, Pennsylvania, USA). These fasting measures of insulin sensitivity are well correlated with the euglycemic hyperinsulinemic clamp [18].

Assessment of exposure to antiretroviral therapy and covariates

Each semiannual MACS study visit includes questions about the use of specific ART. The definition of highly active antiretroviral therapy (HAART) follows the DHHS/Kaiser Panel guidelines [20] and is (a) two or more NRTIs in combination with at least one PI or one non-nucleoside analogue reverse transcriptase inhibitor (NNRTI); (b) one NRTI in combination with at least one PI and at least one NNRTIs; (c) a regimen containing ritonavir and saquinavir in combination with one NRTIs and no NNRTIs; or (d) an abacavir- or tenofovir-containing regimen of three or more NRTIs in the absence of both PI and NNRTIs. All other combined ART regimens, including combinations of zidovudine or stavudine with either a PI or NNRTI, were classified as combination therapy.

The participants were classified into five groups based on self-reported ART use over the 6 months prior to each study visit: (a) HIV seronegative, (b) HIV infected with no ART, (c) HIV infected with monotherapy or combination therapy, (d) HIV infected with non-PI-based HAART, and (e) HIV infected with PI-based HAART. Subjects in the monotherapy and combination therapy groups were evaluated together because there were few (n = 21) person-visits in the monotherapy group throughout the study period and the insulin levels in the two groups were very similar.

Cumulative exposure to NRTIs, PIs, and NNRTIs was quantified using three continuous variables denoting years of use for each therapy class and was updated at each visit. Age, body mass index (BMI; weight in kilograms divided by height in meters squared), nadir CD4 cell count, hepatitis C virus antibody serostatus, all measured at the index visit, and race (Caucasian or non-Caucasian) were ascertained for all participants. Hepatitis C was included as a covariate because of its association with hyperglycemia in HIV-infected persons taking HAART [21]. Family history of diabetes was elicited at each semiannual study visit beginning in March 2000. For the 761 person-visits before March 2000, when family history had not been elicited, the first reported family history was assigned.

Statistical analysis

Fisher's exact and Wilcoxon's non-parametric tests were used to examine the differences between HIV-seronegative and HIV-infected men at the index visit. To examine the effects of HIV infection and ART exposure on insulin resistance markers, multiple linear and logistic regression analyses were conducted for the continuously distributed QUICKI and the binary indicator of insulin > 15 μU/ml, respectively. For both the repeated measures of QUICKI and the repeated assessments of insulin, the within-subject dependence was treated as a nuisance parameter by calculating robust variance estimates using generalized estimating equations with an independent working covariance matrix [22]. All point estimates and corresponding 95% confidence intervals (CI) were obtained using SAS software version 8 (SAS Institute Inc, Cary, North Carolina, USA).

For each model, a specification of exposure (see below) was included, as well as a single continuous covariate representing time in years since the index visit. All models were adjusted for age (centered at 48 years), BMI (centered at 26 kg/m2), nadir CD4 cell count, and binary indicators for hepatitis C virus status, race, and family history of diabetes mellitus. Nadir CD4 cell count was transformed to allow meaningful comparisons of the HIV-infected subgroups with the HIV-seronegative group. For HIV-infected men, the nadir CD4 variable was rescaled as (value − 250)/50. The effect of nadir CD4 cell count was not included in the model for HIV-seronegative men.

In the first model, recent exposure was specified by including four time-dependent binary indicators for the five-level categorical HIV/ART variable previously defined, where the HIV-seronegative men were the reference group and this variable reflects the HIV and ART status in the prior 6 months. In a second model, HIV infection was specified by replacing the four time-dependent binary indicators with one indicator of HIV infection, and the cumulative exposure was specified by replacing the three time-dependent continuous variables separately quantifying the cumulative years of reported exposure to NRTIs, PIs, and NNRTIs beginning 3 years before the index visit. The aim in using this cut-off point at 3 years, which corresponds approximately to the time which PI were introduced, was to confine NRTI exposure to the HAART era. Because NRTIs were introduced much earlier, there was concern that the duration of NRTI exposure since the study entry would be a marker for disease duration.

In a third model, specific ART exposures were explored by restricting the analysis to the 434 HIV-infected men used in model 2 and estimating the effect of the five most frequently used PI, the four most frequently used NRTIs, and the two most frequently used NNRTIs, as observed at the index visit. Experience with specific antiretroviral mediations was quantified using cumulative time-dependent exposures. Among the PI drugs, ritonavir exposure was divided into ‘full dose’ or ‘boosting dose’. Ritonavir exposure was considered to be full dose when it was received as the only PI. Ritonavir administration in combination with another PI, including lopinavir, was considered a boosting dose. The relation between each medication exposure and each endpoint was further adjusted for the other two classes of ART, which were also treated as time-dependent cumulative exposures.

When measurements were missing for insulin, glucose or ART use (1113 of 5734, or 19% of the expected data points), results were carried forward from the most recent prior value. A group of 259 men (who contributed 626 person-visits) who were missing any covariate data was excluded from multiple regression models. This included 52 men missing BMI assessment at the index visit (another 51 had values carried forward from a prior value within 2 years); 199 men missing hepatitis C virus serostatus determinations; four men missing ART information at all of their visits; and four men missing family history records of diabetes mellitus.

Results

Study population

Compared with the 755 HIV-seronegative men, the 533 HIV-infected men were younger and had slightly lower BMI values, higher fasting glucose and insulin concentrations, and lower QUICKI values at the index visit (Table 1). Excluding visits with missing data, these 1288 men contributed 5105 person-visits, which were used in multiple regression analyses, including 2735 HIV-seronegative person-visits. For the HIV-infected men, there were 426 non-ART person-visits, 14 monotherapy person-visits, 156 combination therapy person-visits, 604 non-PI-HAART person-visits, and 1170 PI-HAART person-visits.

Table 1
Table 1:
Characteristics of 1288 men at the index visit.

Effect of recent exposure to antiretroviral therapy by groups

When the QUICKI index was used as a measure of insulin sensitivity, all of the HIV-infected subgroups defined by therapy had significantly lower values (i.e., more insulin resistance) than the HIV-seronegative group. The largest difference was observed in the HIV-infected men who had received PI-based HAART in the previous 6 months (Fig. 1), in a multivariate model. However, this difference in QUICKI was not statistically different from the non-PI-HAART group (P = 0.369) or the mono/combination therapy group (P = 0.127); it was statistically different from the no ART group (P = 0.003). Other factors in the multivariate model independently associated with a lower QUICKI value were age (per 5 years’ increase: mean difference, −0.08; 95% CI, −0.12 to −0.03), BMI (per 5 units increase: mean difference, −0.53; 95% CI, −0.62 to −0.44), and nadir CD4 cell count (per 50 × 106 cells/l decrease, mean difference, −0.05; 95% CI, −0.08 to −0.01).

Fig. 1
Fig. 1:
Mean differences in QUICKI and odds ratios for increased insulin for groups of HIV-positive individuals relative to the HIV-negative group. Therapy in HIV-positive group was defined by treatment received in the 6 months prior to assessment (non-ART, no antiretroviral therapy; mono/combo, monotherapy or non-HAART combination therapy; non-PI-HAART, HAART regimen without a protease inhibitor; PI-HAART, HAART regimen including a protease inhibitor). The vertical bars represent 95% confidence intervals for the point estimates. All regression models were adjusted for age, body mass index, race, nadir CD4 cell count, hepatitis C serostatus and family history of diabetes mellitus. QUICKI, index of insulin sensitivity (see Methods); HAART, highly active antiretroviral therapy.

Similarly, the likelihood of fasting hyperinsulinemia was also increased in each of the HIV-infected treatment groups relative to the HIV-seronegative group, with the greatest difference in the group that received PI-HAART (Fig. 1). Additional factors associated with hyperinsulinemia were similar to those seen in the QUICKI model, with the exception that the effect of nadir CD4 cell count was no longer significant (P = 0.09). Non-Caucasian men had an odds ratio (OR) of 1.5 for elevated insulin compared with Caucasian men (95% CI, 1.1–2.1). In both the QUICKI and the hyperinsulinemia models, hepatitis C virus seropositivity [mean difference, −0.18 (95% CI, −0.51 to 0.14); OR, 1.7 (95% CI, 0.7–2.2)] and family history of diabetes [mean difference, −0.01 (95% CI, −0.15 to 0.12); OR, 1.3 (95% CI, 0.9–1.6)] were not associated with insulin resistance markers.

Effect of cumulative exposure to antiretroviral therapy by drug class

Of the three major ART classes (Table 2), only cumulative exposure to NRTI was significantly and independently associated with lower QUICKI values and increased odds of fasting hyperinsulinemia. Cumulative exposure to PI tended to be associated with fasting hyperinsulinemia (for each additional year of exposure: OR,1.06; 95% CI, 0.99–1.14), but this result was not statistically significant (P = 0.08). Cumulative exposure to NNRTI was not associated with either marker of insulin resistance. As in the previous model, increasing age and BMI were each associated with lower QUICKI values (i.e., increased insulin resistance) and fasting hyperinsulinemia. A lower nadir CD4 cell count was related to a lower QUICKI value and increased odds of fasting hyperinsulinemia when medication class use was treated as a cumulative exposure. Also similar to the first model, non-Caucasian race was associated with increased odds of hyperinsulinemia, but not a lower QUICKI index. Changing the reference group to HIV-infected men without ART exposure yielded similar results (data not shown).

Table 2
Table 2:
The effect of cumulative drug exposure (counted from 3 years prior to index visit) by class and other factors on surrogate insulin resistance markers.

Effect of cumulative exposure to specific antiretroviral medications

The most commonly used antiretroviral medications at the index visit in each class are shown in Table 1. Of these, stavudine and lamivudine exposures were associated with significantly lower QUICKI values (i.e., increased insulin resistance) and increased OR of fasting hyperinsulinemia with each year of use, independent of PI or NNRTI use (Table 3). Of the PI drugs investigated, only indinavir use was associated with increased odds of hyperinsulinemia (per additional year of use: OR, 1.14; 95% CI 1.02–1.26), but it was not associated with a difference in QUICKI index. Cumulative exposure to either efavirenz or nevirapine was not associated with either marker of insulin resistance. Treating medication exposures as ‘ever or never’ yielded a similar pattern of results (data not shown).

Table 3
Table 3:
The effect of cumulative exposure to specific antiretroviral medications on surrogate markers of insulin resistance among HIV-infected subjects.

Discussion

In this study of a large group of HIV-infected men on a wide variety of ART regimens, the presence of HIV infection regardless of ART treatment status was associated with decreased insulin sensitivity as measured by the QUICKI index, as well as an increased risk of fasting hyperinsulinemia, when compared with the HIV-seronegative men. Cumulative exposure NRTI, but not PI, was independently associated with these fasting markers of insulin resistance.

As a common component of all HAART regimens, NRTI have been implicated in many of the metabolic and morphological abnormalities observed in treated HIV-infected patients, particularly peripheral fat wasting [23]. Several studies have shown an association between abnormalities in glucose metabolism and NRTI exposure. Hadigan et al. [1] have previously reported on the relationship between insulin resistance and NRTI exposure in a subanalysis of a case–control study that included 71 HIV-infected patients. In that study, duration of NRTI exposure was the only significant predictor of fasting hyperinsulinemia (1.6 μU/ml for each additional year; P = 0.02) independent of age, gender, waist–hip ratio, BMI, current PI use, total duration of PI use and total duration of NRTI use. Supporting this association, Brambilla et al. [24] found that stavudine exposure was associated with incident diabetes mellitus in an Italian cohort. In our study, we found that cumulative exposure to NRTI, particularly lamivudine and stavudine, had the strongest association of any drug class with fasting markers of insulin resistance.

The mechanisms underlying this association have not been clarified. Exposure to NRTI has been strongly implicated in the peripheral fat wasting observed in HIV-infected patients taking HAART [23], likely through the inhibition of mitochondrial DNA polymerase-gamma [25]. Fat atrophy induced by NRTIs also may lead to the elaboration of mediators important in the pathogenesis of insulin resistance. Apoptotic adipocytes release free fatty acids, which accumulate in other organs, including the liver and skeletal muscle, leading to hepatic [26] and peripheral insulin resistance [3], in a process that has been termed ‘systemic steatosis’ [27]. In addition, abnormalities in adipocyte metabolism induced by NRTIs may lead to increased circulating levels of adipocytokines, such as tumor necrosis factor-α and interleukin-6 [28], causing reduced peripheral insulin sensitivity.

In previous studies, exposure to PI has been associated with insulin resistance. Evidence for their effect on insulin sensitivity has come from ‘switch studies’ demonstrating improvement in glucose metabolism after discontinuation of PIs [29], healthy volunteer studies showing reduced insulin sensitivity after short-term administration of indinavir [8,30], and in vitro studies showing direct effects of certain PI drugs on glucose movement through the GLUT4 transporter [31]. In our study, those HIV-infected patients taking PI-containing HAART regimens in the 6 months prior to assessment appeared to have the most insulin resistance. However, cumulative exposure to PI drugs as a class showed no increased risk of elevated insulin resistance markers.

A number of explanations for the relatively modest effects of PI drugs are possible. First, the effect of PI on insulin sensitivity is heterogeneous [12], and examining PI drugs as a class may have diluted any effect of individual medications. In the present study, cumulative indinavir exposure, but not other PIs, was associated with fasting hyperinsulinemia. Second, some PIs have been shown to impair beta cell function [32]. For this reason, an insulin resistance marker that relies on the compensatory hyperinsulinemic response to hyperglycemia may not accurately reflect peripheral sensitivity to insulin.

In addition to ART exposure, other factors, including those related to the host and to HIV itself, are likely important in the pathogenesis of insulin resistance in HIV-infected patients. In our study, a lower QUICKI index and increased odds of fasting hyperinsulinemia were also observed among those HIV-infected men not exposed to ART in the 6 months prior to assessment. While some of the effect may have been a residual consequence of prior medication exposure, an effect of HIV infection itself may explain these findings. It has been postulated that chronic HIV infection and the resulting cytokine elaboration may contribute to hyperglycemia [33].

Nadir CD4 cell count has been associated with body composition changes in HIV-infected patients [34,35]. This is the first report to our knowledge that links HIV disease severity with abnormalities in glucose homeostasis. The nature of the relation between disease severity and the pathogenesis of metabolic and morphological abnormalities has not been established but may be related to disordered cytokine expression in T cells after immune reconstitution [36].

To estimate insulin sensitivity, we used the both the QUICKI index and a fasting insulin concentration. In the general population, the QUICKI index has been shown to correlate well with measures derived from the clamp technique [18,37,38] and has proven to be a powerful predictor of incident diabetes mellitus [39,40]. There are, however, important limitations to fasting surrogate markers of insulin sensitivity that should be recognized. Correlations with standard methodologies, such as the insulin clamp, appear to depend on the population studied, with stronger associations demonstrable in obese and diabetic persons [18,41]. In HIV-infected persons, QUICKI and other fasting measures of insulin sensitivity were significantly, yet modestly (r2 = 0.18–0.37), correlated with a reference technique comparable to the insulin clamp [42]. In contrast to reference methodologies, fasting measurements of glucose and insulin reflect other aspects of glucose homeostasis, such as insulin secretion and hepatic glucose production, in addition to the effect of insulin on peripheral tissue.

Nevertheless, although seemingly modest, the observed differences in fasting surrogate measures may be clinically important because of their potential association with diabetes and atherosclerosis. In our study, men exposed to PI-containing HAART in the 6 months prior to evaluation had a 5% lower QUICKI value compared with the HIV-seronegative men. In a large Finnish study of HIV-seronegative obese persons [39], a difference in QUICKI (between the first and the third tertiles) of 6% was associated with an increase in the risk developing type 2 diabetes over 5 years of follow-up (OR, 7.8; 95% CI, 1.6–37). In another study of HIV-seronegative participants, the adjusted OR of severe carotid atherosclerosis was 5.7 times higher in those participants in the highest tertile of QUICKI scores compared with the lowest two tertiles, even though the crude difference in QUICKI scores was only 5.5% [43].

Our study had several other limitations. First, the MACS is comprised only of men. Whether these findings are generalizable to women is unknown. Second, men who were exposed to medications affecting serum insulin concentrations, including antidiabetic medications, were excluded. This may have led to an underestimation of the relative differences in insulin resistance markers, since the study population was likely depleted of some subjects with insulin resistance. In addition, because we restricted cumulative exposure to the HAART era, cumulative exposure to NRTIs may have been also underestimated, as these drugs were introduced much earlier. Next, the timing of the glucose and insulin measurements was not standardized in relation to the administration of antiretroviral medications. Because PI can have direct and immediate effects on insulin resistance, an acute and transient antiretroviral effect cannot be ruled out. Further, although we attempted to adjust for potential confounding factors, these observational data may be confounded by unmeasured variables strongly associated with both ART and insulin resistance. Finally, medication histories relied on self-report.

In conclusion, cumulative exposure to NRTIs, but not PIs or NNRTIs, was independently associated with two markers of insulin resistance. These findings highlight the fact that insulin resistance in HIV-infected persons is attributable to other factors besides the use of PIs, including disease-related factors. The extent to which insulin resistance contributes to an increased risk of diabetes mellitus and cardiovascular disease in HIV-infected persons remains to be determined.

Sponsorship: This work was supported by the National Institute of Allergy and Infectious Diseases, with additional supplemental funding from the National Cancer Institute: U01–AI-35042, 5–M01–RR-00052 (GCRC), U01–AI-35043, U01–AI-37984, U01–AI-35039, U01–AI-35040, U01–AI-37613, U01–AI-35041.

Note: The National Institute of Allergy and Infectious Diseases and the National Cancer Institute had representatives on the MACS Executive Committee that oversaw the management of the study and the data collection. The sponsors had no role in the analyses, manuscript preparation or authorization for publication.

References

1. Hadigan C, Meigs JB, Corcoran C, Rietschel P, Piecuch S, Basgoz N, 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.
2. Walli R, Herfort O, Michl GM, Demant T, Jager H, Dieterle C, et al. Treatment with protease inhibitors associated with peripheral insulin resistance and impaired oral glucose tolerance in HIV-1-infected patients. AIDS 1998; 12:F167–F173.
3. Gan SK, Samaras K, Thompson CH, Kraegen EW, Carr A, Cooper DA, et al. Altered myocellular and abdominal fat partitioning predict disturbance in insulin action in HIV protease inhibitor-related lipodystrophy. Diabetes 2002; 51:3163–3169.
4. Mynarcik DC, McNurlan MA, Steigbigel RT, Fuhrer J, Gelato MC. Association of severe insulin resistance with both loss of limb fat and elevated serum tumor necrosis factor receptor levels in HIV lipodystrophy. J Acquir Immune Defic Syndr 2000; 25:312–321.
5. Mynarcik DC, Combs T, McNurlan MA, Scherer PE, Komaroff E, Gelato MC. Adiponectin and leptin levels in HIV-infected subjects with insulin resistance and body fat redistribution. J Acquir Immune Defic Syndr 2002; 31:514–520.
6. Vigouroux C, Maachi M, Nguyen TH, Coussieu C, Gharakhanian S, Funahashi T, et al. Serum adipocytokines are related to lipodystrophy and metabolic disorders in HIV-infected men under antiretroviral therapy. AIDS 2003; 17:1503–1511.
7. Bastard JP, Caron M, Vidal H, Jan V, Auclair M, Vigouroux C, et al. Association between altered expression of adipogenic factor SREBP1 in lipoatrophic adipose tissue from HIV-1-infected patients and abnormal adipocyte differentiation and insulin resistance. Lancet 2002; 359:1026–1031.
8. Noor MA, Seneviratne T, Aweeka FT, Lo JC, Schwarz JM, Mulligan K, et al. Indinavir acutely inhibits insulin-stimulated glucose disposal in humans: a randomized, placebo-controlled study. AIDS 2002; 16:F1–F8.
9. Noor MA, Parker RA, O'Mara E, Grasela DM, Currie A, Hodder SL, et al. The effects of HIV protease inhibitors atazanavir and lopinavir/ritonavir on insulin sensitivity in HIV-seronegative healthy adults. AIDS 2004; 18:2137–2144.
10. Nightingale SL. From the Food and Drug Administration. JAMA 1997; 278:379.
11. Woerle HJ, Mariuz PR, Meyer C, Reichman RC, Popa EM, Dostou JM, et al. Mechanisms for the deterioration in glucose tolerance associated with HIV protease inhibitor regimens. Diabetes 2003; 52:918–925.
12. Dubé MP, Qian D, Edmondson-Melancon H, Sattler FR, Goodwin D, Martinez C, et al. Prospective, intensive study of metabolic changes associated with 48 weeks of amprenavir-based antiretroviral therapy. Clin Infect Dis 2002; 35:475–481.
13. Hadigan C, Miller K, Corcoran C, Anderson E, Basgoz N, Grinspoon S. Fasting hyperinsulinemia and changes in regional body composition in human immunodeficiency virus-infected women. J Clin Endocrinol Metab 1999; 84:1932–1937.
14. Hadigan C, Corcoran C, Stanley T, Piecuch S, Klibanski A, Grinspoon S. Fasting hyperinsulinemia in human immunodeficiency virus-infected men: relationship to body composition, gonadal function, and protease inhibitor use. J Clin Endocrinol Metab 2000; 85:35–41.
15. Meininger G, Hadigan C, Rietschel P, Grinspoon S. Body-composition measurements as predictors of glucose and insulin abnormalities in HIV-positive men. Am J Clin Nutr 2002; 76:460–465.
16. Vigouroux C, Gharakhanian S, Salhi Y, Nguyen TH, Chevenne D, Capeau J, et al. Diabetes, insulin resistance and dyslipidaemia in lipodystrophic HIV-infected patients on highly active antiretroviral therapy (HAART). Diabetes Metab 1999; 25:225–232.
17. Kaslow RA, Ostrow DG, Detels R, Phair JP, Polk BF, Rinaldo CR Jr. The Multicenter AIDS Cohort Study: rationale, organization, and selected characteristics of the participants. Am J Epidemiol 1987; 126:310–318.
18. Katz A, Nambi SS, Mather K, Baron AD, Follmann DA, Sullivan G, et al. Quantitative insulin sensitivity check index: a simple, accurate method for assessing insulin sensitivity in humans. J Clin Endocrinol Metab 2000; 85:2402–2410.
19. Bondar RJ, Mead DC. Evaluation of glucose-6-phosphate dehydrogenase from Leuconostoc mesenteroides in the hexokinase method for determining glucose in serum. Clin Chem 1974; 20:586–590.
20. Dybul M, Fauci AS, Bartlett JG, Kaplan JE, Pau AK. Guidelines for using antiretroviral agents among HIV-infected adults and adolescents. Ann Intern Med 2002; 137:381–433.
21. Mehta SH, Moore RD, Thomas DL, Chaisson RE, Sulkowski MS. The effect of HAART and HCV infection on the development of hyperglycemia among HIV-infected persons. J Acquir Immune Defic Syndr 2003; 33:577–584.
22. Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986; 42:121–130.
23. Mallal SA, John M, Moore CB, James IR, McKinnon EJ. Contribution of nucleoside analogue reverse transcriptase inhibitors to subcutaneous fat wasting in patients with HIV infection. AIDS 2000; 14:1309–1316.
24. Brambilla AM, Novati R, Calori G, Meneghini E, Vacchini D, Luzi L, et al. Stavudine or indinavir-containing regimens are associated with an increased risk of diabetes mellitus in HIV-infected individuals. AIDS 2003; 17:1993–1995.
25. Brinkman K, Smeitink JA, Romijn JA, Reiss P. Mitochondrial toxicity induced by nucleoside-analogue reverse-transcriptase inhibitors is a key factor in the pathogenesis of antiretroviral-therapy-related lipodystrophy. Lancet 1999; 354:1112–1115.
26. Sutinen J, Hakkinen AM, Westerbacka J, Seppala-Lindroos A, Vehkavaara S, Halavaara J, et al. Increased fat accumulation in the liver in HIV-infected patients with antiretroviral therapy-associated lipodystrophy. AIDS 2002; 16:2183–2193.
27. Balasubramanyam A, Sekhar RV, Jahoor F, Jones PH, Pownall HJ. Pathophysiology of dyslipidemia and increased cardiovascular risk in HIV lipodystrophy: a model of ‘systemic steatosis’. Curr Opin Lipidol 2004; 15:59–67.
28. Mynarcik DC, McNurlan MA, Steigbigel RT, Fuhrer J, Gelato MC. Association of severe insulin resistance with both loss of limb fat and elevated serum tumor necrosis factor receptor levels in HIV lipodystrophy. J Acquir Immune Defic Syndr 2000; 25:312–321.
29. Walli RK, Michl GM, Bogner JR, Goebel FD. Improvement of HAART-associated insulin resistance and dyslipidemia after replacement of protease inhibitors with abacavir. Eur J Med Res 2001; 6:413–421.
30. Noor MA, Lo JC, Mulligan K, Schwarz JM, Halvorsen RA, Schambelan M, et al. Metabolic effects of indinavir in healthy HIV-seronegative men. AIDS 2001; 15:F11–F18.
31. Murata H, Hruz PW, Mueckler M. The mechanism of insulin resistance caused by HIV protease inhibitor therapy. J Biol Chem 2000; 275:20251–20254.
32. Dufer M, Neye Y, Krippeit-Drews P, Drews G. Direct interference of HIV protease inhibitors with pancreatic beta-cell function.Naunyn Schmiedebergs Arch Pharmacol 2004.
33. Gelato MC. Insulin and carbohydrate dysregulation. Clin Infect Dis 2003; 36:S91–S95.
34. Lichtenstein KA, Delaney KM, Armon C, Ward DJ, Moorman AC, Wood KC, et al. Incidence of and risk factors for lipoatrophy (abnormal fat loss) in ambulatory HIV-1-infected patients. J Acquir Immune Defic Syndr 2003; 32:48–56.
35. Mauss S, Corzillius M, Wolf E, Schwenk A, Adam A, Jaeger H, et al. Risk factors for the HIV-associated lipodystrophy syndrome in a closed cohort of patients after 3 years of antiretroviral treatment. HIV Med 2002; 3:49–55.
36. Ledru E, Christeff N, Patey O, de Truchis P, Melchior JC, Gougeon ML. Alteration of tumor necrosis factor-alpha T-cell homeostasis following potent antiretroviral therapy: contribution to the development of human immunodeficiency virus-associated lipodystrophy syndrome. Blood 2000; 95:3191–3198.
37. Yokoyama H, Emoto M, Fujiwara S, Motoyama K, Morioka T, Komatsu M, et al. Quantitative insulin sensitivity check index and the reciprocal index of homeostasis model assessment are useful indexes of insulin resistance in type 2 diabetic patients with wide range of fasting plasma glucose. J Clin Endocrinol Metab 2004; 89:1481–1484.
38. Rabasa-Lhoret R, Bastard JP, Jan V, Ducluzeau PH, Andreelli F, Guebre F, et al. Modified quantitative insulin sensitivity check index is better correlated to hyperinsulinemic glucose clamp than other fasting-based index of insulin sensitivity in different insulin-resistant states. J Clin Endocrinol Metab 2003; 88:4917–4923.
39. Vanhala P, Vanhala M, Kumpusalo E, Keinanen-Kiukaanniemi S. The quantitative insulin sensitivity check index QUICKI predicts the onset of type 2 diabetes better than fasting plasma insulin in obese subjects: a 5-year follow-up study. J Clin Endocrinol Metab 2002; 87:5834–5837.
40. Hanley AJ, Williams K, Gonzalez C, D'Agostino RB Jr, Wagenknecht LE, Stern MP, et al. Prediction of type 2 diabetes using simple measures of insulin resistance: combined results from the San Antonio Heart Study, the Mexico City Diabetes Study, and the Insulin Resistance Atherosclerosis Study. Diabetes 2003; 52:463–469.
41. Mather KJ, Hunt AE, Steinberg HO, Paradisi G, Hook G, Katz A, et al. Repeatability characteristics of simple indices of insulin resistance: implications for research applications. J Clin Endocrinol Metab 2001; 86:5457–5464.
42. Chu JW, Abbasi F, Beatty GW, Khalili M, Koch J, Rosen A, et al. Methods for quantifying insulin resistance in human immunodeficiency virus-positive patients. Metabolism 2003; 52:858–861.
43. Rajala U, Laakso M, Paivansalo M, Pelkonen O, Suramo I, Keinanen-Kiukaanniemi S. Low insulin sensitivity measured by both quantitative insulin sensitivity check index and homeostasis model assessment method as a risk factor of increased intima-media thickness of the carotid artery. J Clin Endocrinol Metab 2002; 87:5092–5097.

Appendix

Members of MACS (http://www.statepi.jhsph.edu/macs/macs.html): Joseph B. Margolick (Principal Investigator), Haroutune Armenian, Adrian Dobs, Homayoon Farzadegan, Shenghan Lai, Justin McArthur, Chloe Thio (The Johns Hopkins University Bloomberg School of Public Health: Baltimore); John P. Phair (Principal Investigator), Joan S. Chmiel (Co-Principal Investigator), Sheila Badri, Bruce Cohen, Craig Conover, Maurice O'Gorman, Frank Pallela, Daina Variakojis, Steven M. Wolinsky (Howard Brown Health Center, Feinberg School of Medicine, Northwestern University, and Cook County Bureau of Health Services, Chicago); Roger Detels and Beth Jamieson (Principal Investigators), Barbara R. Visscher (Co-Principal Investigator), Anthony Butch, John Fahey, Otoniel Martínez-Maza, Eric N. Miller, John Oishi, Paul Satz, Elyse Singer, Harry Vinters, Otto Yang, Stephen Young (Schools of Public Health and Medicine, University of California at Los Angeles, Los Angeles); Charles R. Rinaldo (Principal Investigator), Lawrence Kingsley (Co-Principal Investigator), James T. Becker, Phalguni Gupta, John Mellors, Sharon Riddler, Anthony Silvestre (University of Pittsburgh Graduate School of Public Health, Pittsburgh).

Data Coordinating Center at The Johns Hopkins University Bloomberg School of Public Health: Lisa P. Jacobson (Principal Investigator), Haitao Chu, Stephen R. Cole, Xiuhong Li, Alvaro Muñoz, Janet Schollenberger, Eric Seaberg, Michael Silverberg, Sol Su.

National Institute of Allergy and Infectious Diseases: Robin Huebner (National Cancer Institute), Jodi Black.

Keywords:

antiretroviral therapy; insulin resistance; HIV; glucose; hyperinsulinemia; protease inhibitors; nucleoside analogue reverse transcriptase inhibitors

© 2005 Lippincott Williams & Wilkins, Inc.