Share this article on:

Mixed Patterns of Changes in Central and Peripheral Fat Following Initiation of Antiretroviral Therapy in a Randomized Trial

Mulligan, Kathleen PhD*; Parker, Robert A. ScD; Komarow, Lauren MS; Grinspoon, Steven K. MD; Tebas, Pablo MD§; Robbins, Gregory K. MD; Roubenoff, Ronenn MD, MHS; Dubé, Michael P. MDfor the ACTG 384 and A5005s Study Teams

JAIDS Journal of Acquired Immune Deficiency Syndromes: April 15th, 2006 - Volume 41 - Issue 5 - p 590-597
doi: 10.1097/01.qai.0000214811.72916.67
Clinical Science

Introduction: Although peripheral fat loss is a consistent feature of HIV-associated lipodystrophy, less is known about whether changes in central and peripheral fat occur in tandem or independently over time. We examined intraindividual changes in dual-energy x-ray absorptiometry (DEXA) and anthropometry data over 64 weeks after initiation of antiretroviral therapy (ART).

Methods: In ACTG 384, ART-naive subjects were randomized to receive didanosine and stavudine or zidovudine and lamivudine plus efavirenz, nelfinavir, or both. In the metabolic substudy A5005S (N'= 329), waist, hip, thigh, and arm circumferences were measured in triplicate. DEXA scanning was performed in 157 subjects.

Results: Individual changes in limb and trunk fat at week 64 correlated positively (R =.72, P <.001). Most subjects had directionally concordant changes in limb and trunk fat (36% gained and 32% lost in both regions). In 26%, trunk fat increased while limb fat decreased, whereas the reverse occurred in only 6% (P <.001). Changes in waist and hip circumferences also correlated positively (R =.62, P <.001). The proportion of subjects with waist/hip ratio more than.95 (men) or.90 (women) was 34% at baseline and 47% at week 64 (P =.003). In those who developed increased waist/hip ratio, 76% had increased waist circumference.

Conclusions: Individual results obtained by both anthropometry and DEXA show diverse patterns of fat gain and loss over 64 weeks after initiation of ART, but changes tended to occur in the same direction. Increased waist/hip ratio was predominantly associated with increased waist circumference. There was no evidence that central fat accumulation and peripheral fat loss are linked in most subjects.

From the *Department of Medicine, Division of Endocrinology, San Francisco General Hospital, University of California at San Francisco, San Francisco, CA; †Statistical and Data Analysis Center, Harvard University, Boston, MA; ‡Massachusetts General Hospital, Harvard University, Boston, MA; §University of Pennsylvania, Philadelphia, PA; ∥Friedman School of Nutrition Science and Policy, Tufts University, Boston, MA; and ¶Indiana University, Indianapolis, IN.

Received for publication October 2, 2005; accepted January 18, 2006.

Supported by grants to the AIDS Clinical Trials Group from the National Institute of Allergy and Infectious Diseases (AI38855 and AI38858). The parent protocol (ACTG 384) was supported in part by Agouron Pharmaceuticals, Inc., Bristol-Myers Squibb Company, DuPont Pharmaceutical Company, GlaxoSmithKline, Inc., and Merck and Co., Inc.

These data were presented in part at the 12th Conference on Retroviruses and Opportunistic Infections, February 2005, Boston, MA.

Reprints: Michael P. Dubé, MD, Wishard Memorial Hospital, 1001 W. 10th Street, Suite OPW-430, Indianapolis, IN 46202 (e-mail:

Since the introduction of highly active antiretroviral therapy, a variety of morphological alterations have been recognized as complications of therapy. Some early studies reported the occurrence of peripheral fat loss or lipoatrophy,1,2 whereas others focused on central fat accumulation.3-6 Collectively, these changes have come to be commonly referred to as the HIV lipodystrophy syndrome. Largely on the basis of results of cross-sectional studies in which group values for peripheral fat were lower in HIV-infected patients on antiretroviral therapy, when compared with seronegative or untreated HIV-positive controls,1,7-9 lipoatrophy has come to be regarded as the defining morphological feature of the lipodystrophy syndrome.

Longitudinal studies of the etiology of alterations in fat distribution have focused primarily on the effects of certain antiretroviral agents.10-13 However, it is still not known whether increases in central fat and decreases in peripheral fat occur in tandem or independently. Likewise, it is not known whether subjects who develop an elevated waist/hip ratio, which is associated with increased risk of cardiovascular disease in studies in seronegative populations,14,15 do so as a result of increasing waist or decreasing hip circumferences. Accordingly, we examined intraindividual in fat distribution over 64 weeks after initiation of highly active antiretroviral therapy in subjects enrolled in A5005S, the metabolic substudy of ACTG 384.

Back to Top | Article Outline



In ACTG 384, 980 HIV-infected subjects were randomized to receive either zidovudine and lamivudine or didanosine and stavudine in combination with nelfinavir, efavirenz, or both.16,17 Subjects were eligible for this study if they had less than 7 days prior antiretroviral experience and plasma HIV-1 levels >500 copies/mL. In A5005s, a substudy of morphologic and metabolic changes, a total of 334 subjects were enrolled at 23 participating sites in the United States from 1998 to 1999, with follow-up through 2001. Of these, 329 had anthropometric measurements and/or underwent dual-energy x-ray absorptiometry (DEXA) scanning and are included in this report. The study was approved by each study site s institutional review board, and written informed consent was obtained from each subject before enrollment. The primary results of A5005s, which focused on the effects of different antiretroviral treatments on glucose and lipid metabolism and regional fat distribution by DEXA, have been reported.10

Back to Top | Article Outline

Body Composition Measurements

Body composition was assessed at baseline and at 16-week intervals following standardized procedures. Subjects were weighed on calibrated scales after removing shoes, outdoor clothing, and other heavy items. Circumferences of the waist, hip, arm, and thigh were measured to the nearest millimeter using anatomical landmarks as defined for the Third National Health and Nutrition Evaluation Survey.18 Study coordinators underwent central supervised training in these measurements at the beginning of the study. All measurements were made intriplicate, and the median value was used for this analysis. An elevated waist/hip ratio was defined as more than 0.95 and 0.90 in men and women, respectively. These values correspond to the upper tertile of control values in a large multinational study of risk factors for myocardial infarction and were associated with significantly increased risk.15 The criterion used for change in a circumference measurement was an increase or decrease of at least 2.54 cm (1 in).

Whole-body DEXA scanning was performed in a subgroup (N = 157) of subjects at 18 clinical sites.10 All scans were analyzed centrally at Tufts University without knowledge of treatment assignment. A standard phantom was circulated to all sites and scanned on the DEXA instruments used in the study. The same scanner was used for all evaluations in any individual subject whenever possible; when scanners were changed, a protocol was employed to obtain serial measures of the phantom on both old and new machines to obtain an estimate of the instrument-specific variability. Limb fat was the sum of arm and leg fat mass.

Back to Top | Article Outline

Statistical Analyses

Continuous variables are summarized as median (interquartile range), whereas categorical variables are summarized as number and percent of total. For continuous variables, the Wilcoxon signed rank test was used to assess changes over time within a single group, and the Wilcoxon rank sum test was used to compare differences between groups. Spearman rank correlation was used to test the'association between variables. Comparisons between Spearman rank correlations used an approach suggested by Fieller et'al.19 For categorical variables, significance between groups was assessed using Fisher exact test (or extensions), whereas McNemar test was used to test concordance within a group.

Back to Top | Article Outline


Table 1 shows the baseline data on the total population, as well as those with only anthropometric data and those with both anthropometric and DEXA data available. Aside from thigh circumference, which was slightly but significantly lower in subjects for whom DEXA data are available, there were no statistically significant differences between groups among the 13 variables that could be compared.



At baseline, all of the anthropometric measures correlated with weight and total and regional DEXA measurements (P <.001 in each case; Table 2). Changes in circumferences at week 64 correlated moderately with changes in fat measured by DEXA and slightly more strongly with changes in weight (Table 3). All such correlations were positive and highly statistically significant (P <.001). The median change in lean body mass at week 64 was relatively modest (+1.15 kg), and correlations between changes in circumferences and changes in total lean body mass were of a lesser magnitude than with changes in fat or weight (Table 3).





In the group as a whole, median changes from baseline to week 64 in hip, thigh, and arm circumference were directionally consistent with change in limb fat measured by DEXA, initially increasing but subsequently decreasing (Fig. 1A). There was also directional consistency in the changes in trunk fat by DEXA and waist circumference (Fig. 1B). In both cases, median values increased initially and then plateaued or declined only slightly.



Individual changes in trunk and limb fat at week 64 measured by DEXA correlated positively and highly significantly (R =.72; P <.001; Fig. 2). Changes were directionally concordant in most subjects, with increases in both trunk and limb fat seen in 36% and decreases in both trunk and limb fat in 32%. Discordant changes were noted in one third of subjects: overall, 26% had an increase in trunk fat and a decrease in limb fat, whereas only 6% had a decrease in trunk fat and an increase in limb fat (P <.001, McNemar test). Changes were related to the baseline values: those who gained limb fat started with lower values of limb fat than those who lost limb fat (P = 0.007), and those who gained trunk fat started with lower values of trunk fat than those who lost trunk fat (P =.03). These relationships were independent of whether the pattern was concordant or discordant.



Consistent with the DEXA results, changes in waist and hip circumference at week 64 also correlated positively (R =.62; P <.001; Fig. 3). The proportion of subjects who met criteria for an elevated waist/hip ratio (>.95 and.90 in men and women, respectively) increased from 34% at baseline to 47% at week 64 (P =.003, McNemar test). Fifty subjects developed an elevated waist/hip ratio between baseline and week 64 (shown as closed circles in Fig. 3). In 52% of these subjects, the elevation was a result of an increase in waist circumference with no change in hip circumference; among the others, 24% had both an increase in waist circumference and a decrease in hip circumference, and 22% had a decrease in hip circumference with no increase in waist circumference.



Although subgroup analyses revealed some statistically significant differences in the magnitude of the correlations between men and women and among the 3 different racial/ethnic groups (white/non-Hispanic, black/non-Hispanic, and Hispanic), there were no differences in the direction of any of these correlations. Although there were marked differences between treatment arm assignment (zidovudine and lamivudine or didanosine and stavudine, or nelfinavir, efavirenz, or both) in the changes observed over time,10 the correlations between anthropometric and DEXA measurements were similar across the individual treatment arm assignments. Thus, the observed relationships were not driven by results in one racial/ethnic group or a specific treatment.

Back to Top | Article Outline


In this series of analyses, we examined changes in regional fat measured by DEXA and circumferences over 64 weeks in HIV-infected individuals initiating their first antiretroviral regimen. Median changes using both techniques showed patterns of change that are similar to those in other randomized13 and nonrandomized20 prospective studies of initiation of antiretroviral regimens using similar combinations of NRTIs. Rather than focus on the effects of specific treatment regimens, our approach in this analysis was to examine individual patterns of change in an attempt to determine whether central fat accumulation and peripheral fat loss occurred concurrently or independently, or if one phenomenon preceded or dominated the other.

Analysis of both DEXA and anthropometry results demonstrates that net regional changes over the first 64 weeks were concordant in two thirds of subjects, with approximately equal proportions of subjects gaining or losing in both regions. In three quarters of subjects who developed an elevated waist/hip ratio at week 64, this elevation occurred in conjunction with an increase in waist circumference, with or without a concomitant decrease in hip circumference. In contrast, fewer than 25% of subjects developed an elevated waist/hip ratio as a result of decreased hip circumference alone. Directionally opposite changes in central and peripheral measurements occurred in only a minority of subjects within our period of follow-up, suggesting that central fat accumulation and peripheral fat loss are not linked phenomena. Thus, although net "redistribution" of fat may occur in some individuals, it seems to do so by simultaneous but separate processes. Our longitudinal results are consistent with the findings recently reported in a cross-sectional study of HIV-infected men, most of whom were antiretroviral experienced.7 In that study, there was no association between peripheral fat loss and central fat gain, and there tended to be positive correlations between adipose tissue amounts measured by magnetic resonance imaging in most sites regardless of peripheral or central location.7

It is noteworthy that among the minority of subjects in whom discordant changes occurred, significantly more had net increases in central fat and decreases in peripheral fat than the reverse. Moreover, the plots of both the intraindividual changes in waist circumference versus hip circumference (Fig. 3) and in DEXA trunk fat versus limb fat (Fig. 2) reveal a line of correlation that is shifted downward and to the right. Although changes in central and peripheral fat occurred in the same direction in most subjects, there was an overall greater tendency toward central fat gain rather than loss, and peripheral fat loss rather than gain, over time for the group as a whole. This minority of individuals with increased central fat and decreased peripheral fat may warrant particular scrutiny for the development of adverse changes in glucose and lipid metabolism.21

The physiological significance of an elevated waist/hip ratio that develops as a result of increased waist circumference as opposed to decreased hip circumference is not known. In the general population, an increased waist/hip ratio is usually a result of increased waist circumference. Indeed, waist circumference correlates better than waist/hip ratio with measures of visceral fat and increased cardiovascular risk factors.22,23 In our study, development of an increased waist/hip ratio occurred predominantly in those who experienced increased waist circumference, suggesting that most increases in waist/hip ratio that occur during antiretroviral therapy may be physiologically relevant.24

In this analysis we used results of both DEXA scanning and circumferences. DEXA has been shown to be a sensitive technique for evaluating changes in regional fat distribution and can detect modest but statistically significant changes that are not clinically apparent.25 However, the expense and requirement for specialized equipment can limit the use of DEXA in large trials. We observed moderate but highly statistically significant correlations between changes in total and regional fat by DEXA and all circumferences over 64 weeks. Moreover, although changes in circumferences represent more than changes in fat alone, use of DEXA and circumferences produced results that were directionally consistent for both group trends and changes within individual subjects. We must acknowledge that neither DEXA nor circumferences are able to distinguish between visceral and subcutaneous adipose tissue in the central region, so the current data set does not allow us to speculate on whether the observed increases in waist circumference and trunk fat represent a preferential expansion of visceral as opposed to subcutaneous fat. Another limitation of this study is that we did not have comprehensive diet and exercise information, which may have affected the results.

In summary, we observed a mixed pattern of changes in central and peripheral fat over the first 64 weeks after initiation of antiretroviral therapy. Changes in central and peripheral fat were correlated positively, as were changes in waist and hip circumferences. It was notable that from 32 to 64 weeks, the overall group change in central measures was relatively flat (Fig. 1A), whereas limb fat measures decreased progressively (Fig. 1B). Thus, we cannot exclude the possibility that with longer follow-up, peripheral fat loss may continue and lipoatrophy may emerge as a more dominant feature, as was seen by Mallon et al.20 in a small group of subjects followed through 144 weeks. Nonetheless, these data, collected in a considerably larger group of subjects, suggest that isolated lipoatrophy is not a consistent characteristic of alterations in fat distribution following initiation of antiretroviral therapy.

Back to Top | Article Outline


The authors thank the following A5005s team members who contributed to the design and conduct of the trial: Robert A. Zackin, ScD (SDAC, Harvard University, deceased), Thomas A. Buchanan, MD (University of Southern California), Kevin Yarasheski, PhD (Washington University), Sally Snyder, BS (Social and Scientific Systems), and Jeff Taylor (ACTG Community Constituency Group representative, San Diego, CA). We are grateful to all of the subjects who volunteered for this study. The authors acknowledge the invaluable assistance of Kathy L. Flynn and Gina-Bob Dubé with managing the references, and Abby Shevitz, MD and Jodi L. Forand, BS, of the Friedman School of Nutrition Science and Policy at Tufts University for DEXA scan management.

Back to Top | Article Outline


1. Carr A, Samaras K, Burton S, et al. A syndrome of peripheral lipodystrophy, hyperlipidaemia 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 analogue therapy. AIDS. 1999;13(13):1659-1667.
3. Engelson ES, Kotler DP, Tan YX, et al. Fat distribution in HIV-infected patients reporting truncal enlargement quantified by whole-body magnetic resonance imaging. Am J Clin Nutr. 1999;69:1162-1169.
4. Lo JC, Mulligan K, Tai VW, et al. "Buffalo hump" in men with HIV-1 infection. Lancet. 1998;351:861-870.
5. Miller KD, Jones E, Yanovski JA, et al. Visceral abdominal-fat accumulation associated with use of indinavir. Lancet. 1998; 351:871-875.
6. Miller KK, Daly PA, Sentochnik D, et al. Pseudo-Cushing s syndrome in human immunodeficiency virus-infected patients. Clin Infect Dis. 1998;27:68-72.
7. Bacchetti P, Gripshover B, Grunfeld C, et al. Fat distribution in men with HIV infection. J Acquir Immun Defic Syndr. Oct 1 2005;40(2):121-131.
8. Mulligan K, Tai VW, Algren H, et al. Altered fat distribution in distribution in HIV-positive men on nucleoside analog reverse transcriptase inhibitor therapy. J Acquir Immun Defic Syndr. 2001;26:443-448.
9. 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(1):37-49.
10. Dubé MP, Parker RA, Tebas P, et al. Glucose metabolism, lipid, and body fat changes in antiretroviral-naive subjects randomized to nelfinavir or efavirenz plus dual nucleosides. AIDS. 2005;19:1807-1818.
11. Gallant JE, Staszewski S, Pozniak AL, et al. Efficacy and safety of tenofovir DF vs stavudine in combination therapy in antiretroviral-naive patients: a 3-year randomized trial. JAMA. 2004;292(2):191-201.
12. Joly V, Flandre P, Meiffredy V, et al. Increased risk of lipoatrophy under stavudine in HIV-1-infected patients: results of a substudy from a comparative trial. AIDS. 2002;16(18):2447-2454.
13. Shlay JC, Visnegarwala F, Bartsch G, et al. Body composition and metabolic changes in antiretroviral-naive patients randomized to didanosine and stavudine vs. abacavir and lamivudine. J Acquir Immun Defic Syndr. 2005;38:147-155.
14. Terry RB, Page WF, Haskell WL. Waist/hip ratio, body mass index and premature cardiovascular disease mortality in US Army veterans during a twenty-three year follow-up study. Int J Obes Relat Metab Disord. 1992;16:417-423.
15. Yusuf S, Hawken S, Ounpuu S, et al. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case-control study. Lancet. 2004;364:937-952.
16. Robbins GK, De Gruttola V, Shafer RW, et al. Comparison of sequential three-drug regimens as initial therapy for HIV-1 infection. N Engl J Med. 2003;349(24):2293-2303.
17. Shafer RW, Smeaton LM, Robbins GK, et al. Comparison of four-drug regimens and pairs of sequential three-drug regimens as initial therapy for HIV-1 infection. N Engl J Med. 2003;349(24):2304-2315.
18. Centers for Disease Control. The Third national Health and Nutrition Examination Survey (NHANES III 1988-1994) Reference Manuals and Reports: Body Measurements (Anthropometry). Rockville, MD, 1988.
19. Fieller CE, Hartley HO, Pearson ES. Tests for rank correlation coefficients I. Biometrika. 1957;44:470-481.
20. Mallon PW, Miller J, Cooper DA, et al. Prospective evaluation of the effects of antiretroviral therapy on body composition in HIV-1-infected men starting therapy. AIDS. 2003;17(7):971-979.
21. Vigouroux C, Maachi M, Nguyen TH, et al. Serum adipocytokines are related to lipodystrophy and metabolic disorders in HIV-infected men under antiretroviral therapy. AIDS. 2003;17(10):1503-1511.
22. Janssen I, Katzmarzyk PT, Ross R. Waist circumference and not body mass index explains obesity-related health risk. Am J Clin Nutr. 2004;79:379-384.
23. Pouliot MC, Despres JP, Lemieux S, et al. Waist circumference and abdominal sagittal diameter: best simple anthropometric indexes of abdominal visceral adipose tissue accumulation and related cardiovascular risk in men and women. Am J Cardiol. 1994;73:460-468.
24. Dolan SE, Hadigan C, Killilea KM, et al. Increased cardiovascular disease risk indices in HIV-infected women. J Acquir Immun Defic Syndr. 2005;39:44-54.
25. Carr A, Workman C, Smith DE, et al. Abacavir substitution for nucleoside analogs in patients with HIV lipoatrophy: a randomized trial. JAMA. 2002;288(2):207-215.
Back to Top | Article Outline


The following ACTG investigators and sites participated'in this study: Barbara M. Gripshover, MD and Kathleen Burgner, RN, BSN, Case Western Reserve University (A2501); Ian Frank, MD and Isabel Matozzo, RN, University of Pennsylvania, Philadelphia (A 6201); Laura Laughlin, RN and Diane Gochnour, RN, Ohio State University (A2301) Grant no. U01 AI025924; Tammy Powell, RN, ACRN and Pamposh Kaul, MD, University of Cincinnati (A2401) grant no. AI 25897; Debra deMarco, RN and John Stoneman, RN, Washington University (St. Louis) (A2101) grant nos. AI 25903 and RR-00036; Connie A. Funk, RN, MPH and Kathleen E. Squires, MD, University of Southern California (A1201) grant no. U01 A127673; Margaret A. Fischl, MD and Leslie Thompson, RN, BSN, University of Miami (A0901) grant no. AI27675; Mitch Goldman, MD and Helen Rominger, RN, MSN, FNP, Indiana University Hospital (A2601) grant nos. U01 AI25859 and RR-00750; Linda Meixner, RN and Tari Gilbert, NP, University of California, San Diego (A0701) grant no. AI27670; Christine Fietzer and Kathy A. Fox, University of Minnesota (A1501); Eileen Chusid, PhD and Donna Mildvan, MD (Beth Israel Medical Center), Mount Sinai Medical Center (NY) (A1801) grant no. AI 46370, University of North Carolina (A3201), Howard University (A5301); Lynn Dumas, RN (Beth Israel-Deaconess Hospital), Betsy Adams, RN (Boston Medical Center), and Theresa Flynn, NP (Massachusetts General Hospital), Harvard (A0101); Mallory Witt, MD and Tomasa Maldonado, RN, BS, Harbor General/UCLA School of Medicine (A0601); Juan J.L. Lertora, MD, PhD and Rebecca Clark, MD, PhD, Tulane University (A1701) grant no. IU01AI3844 and GCRC grant no. PHS NCRR M01 RR05096; Pat Cain, RN and Sylvia Stoudt, RN, Stanford University (A0501), NYU/Bellevue (A0401); Harold A. Kessler, MD and Ruth M. Davis, RN, Rush-Presbyterian/St. Lukes (Chicago) (A2702) grant no. UO1 AI025915; Santiago Marrero, MD and Irma Torres, RN, University of Puerto Rico (A5401) grant no. A134832-12; Diane Havlir, MD and Jody Lawrence, MD, San Francisco General Hospital (A0801), University of Hawaii (A5201), University of Rochester Medical Center (A1101); Clifford Gunthel, MD and Ericka Patrick, RN, Emory University (A5802); Michael J. Borucki MD and Gerianne Casey RN, University of Texas, Galveston (A6301) grant no. AI32782; Ilene Wiggins, RN and Dorcas Baker, RN, Johns Hopkins University (A0201) grant nos. RR00052 and AI27668; Carol Dukes-Hamilton, MD and Shelia Tedder, RN, Duke University Medical Center (A1601); Valery Hughes, FNP and Todd Stroberg, RN, Cornell University (A2201) grant no. AI 46386; Sally Canmann, BSN and Cathi Basler, MSN, University of Colorado Health Sciences Center, Denver (A6101); and Beck Royer, PA-C and N. Jeanne Conley, RN, University of Washington (Seattle) (A1401) grant no. AI 27664.


fat distribution; lipodystrophy; DEXA; anthropometry; antiretroviral therapy; HIV

© 2006 Lippincott Williams & Wilkins, Inc.