Recent studies in HIV-infected women suggest that lipoatrophy, particularly in peripheral body sites, seems to be predominant.1-3 Among antiretroviral medications, use of the nucleoside reverse transcriptase inhibitor (NRTI) stavudine has been most strongly associated with peripheral fat loss.1,2 These body fat changes are of major concern because they result in metabolic abnormalities, decreased quality of life, and potentially compromised adherence to otherwise effective antiretroviral therapy (ART).4 One strategy to reverse peripheral fat loss is to discontinue stavudine. Clinical trials in which another NRTI is substituted for stavudine have shown increases in limb fat by dual x-ray absorptiometry (DXA).5-10 These studies have primarily included HIV-infected men, however, and it remains to be seen whether or not the findings can be generalized to HIV-infected women, particularly minorities, who comprise an increasing proportion of HIV-infected patients in the United States.11
Using data from the Women's Interagency HIV Study (WIHS), a cohort that reflects the epidemiology of HIV infection among women, we compare the annual changes in regional anthropometry from July 1999 to March 2005 in HIV-infected women before and after stavudine discontinuation. Because body composition changes also occur with aging, a comparison group of HIV-uninfected women was also studied.
The WIHS is a multicenter prospective cohort study that was established in 1994 to investigate the progression of HIV in women with and at risk for HIV. Of the 3768 participants enrolled at 6 sites (Bronx, NY; Brooklyn, NY; Chicago, IL; Los Angeles, CA; San Francisco, CA; and Washington, DC), 2056 HIV-infected women and 569 HIV-uninfected women were enrolled between October 1994 and November 1995. An additional 737 HIV-infected and 406 HIV-uninfected women were enrolled between October 2001 and September 2002. Sociodemographic characteristics and HIV risk factors were similar between HIV-infected and HIV-uninfected women, and baseline cohort characteristics have been described previously.12,13 Every 6 months, participants receive a comprehensive physical examination and complete an extensive interviewer-administered questionnaire that collects information on demographics, disease characteristics, and specific medication use. The WIHS uses a standard definition of highly active antiretroviral therapy (HAART)14 adapted from the US Department of Health and Human Services/Kaiser Panel guidelines.15 Blood is also collected for CD4 cell count and HIV RNA level. Institutional review boards approved all protocols and consent forms at each of the participating institutions.
Beginning in July 1999, circumferences of the waist, chest, upper arm, hip, and midthigh were collected. Examiners completed a standardized training and certification program conducted by a single trainer before anthropometry data collection. Anthropometry training recertification is conducted every 2 years. Anthropometry measurements were performed using the techniques described by the Third National Health and Examination Survey procedures.16 For each body region, examiners recorded the circumference to the nearest 0.1 cm. Height was recorded in meters and weight in kilograms.
A total of 734 HIV-infected women (82% of all HIV-infected women in the WIHS who reported stavudine use after July 1999) and 698 HIV-uninfected women were included in the analyses because they contributed complete anthropometry data (weight and waist, chest, arm, hip, and thigh circumferences) for at least 2 study visits between July 1999 and March 2005. The 698 HIV-uninfected women included in our analysis satisfied the following criteria. A total of 959 women in the WIHS were HIV-uninfected as of July 1999, 803 of whom had a study visit after July 1999. Six hundred ninety-eight of the 803 women had complete anthropometry data at the same visit at least twice during follow-up. For HIV-infected women, the beginning of the observation period was defined as the first study visit (ie, baseline visit) after July 1999 in which complete anthropometry data were available and stavudine was reported. For HIV-uninfected women, the baseline visit was defined as the first study visit after July 1999 in which complete anthropometry data were available. The 1432 women contributed 8706 study visits (4437 from HIV-infected women and 4269 from HIV-uninfected women), or an average of 6 visits, with complete data on all anthropometry measurements. Follow-up ended at death for 94 participants (81 HIV-infected and 13 HIV-uninfected women), 319 participants (218 HIV-infected and 101 HIV-uninfected women) had their last study visit with complete anthropometry data before April 2004, and the remaining 1019 participants (435 HIV-infected and 584 HIV-uninfected women) completed follow-up with their last study visit with complete anthropometry data between April 2004 and March 2005.
The primary exposures were stavudine discontinuation and HIV status. The reason given most often for discontinuing stavudine was “did not feel like taking any pills,” followed by “to avoid side effects.” Anthropometric changes occurring while on stavudine were compared with changes observed after stavudine discontinuation. These changes were also compared with the changes observed in HIV-uninfected women over the same period. HIV-infected women who discontinued stavudine and then resumed stavudine during follow-up were censored at the study visit before resumption.
The annual change in each anthropometric measure was estimated and compared across the 3 primary exposures using linear regression models with generalized estimating equations to account for the statistical dependence incurred by repeated anthropometric measurements over time on the same individual.17 In our final analyses, individuals discontinuing stavudine were further categorized by the tertiles of the distribution of the number of years since stavudine discontinuation (≤1, 1-2.25, and >2.25 years). All regression models used in the final analyses were adjusted for age; race; prebaseline nadir CD4 cell count; baseline value of the anthropometric measurement that was being modeled; and time-updated values of smoking status, height, and self-reported menopausal status.
Baseline Demographic and Anthropometric Characteristics
Compared with the HIV-uninfected women, HIV-infected women were older and less likely to be African American and more likely to be Hispanic but had similar smoking habits (Table 1). HIV-infected women had lower weight; body mass index (BMI); and circumferences in the chest, arm, hip, and thigh, although waist circumference was comparable.
At the baseline visit, HIV-infected women who reported stavudine use throughout their follow-up (n = 160) and those who later discontinued stavudine (n = 574) were similar with regard to age, race, baseline smoking behavior, CD4 cell count, and body composition (data not shown). Stavudine discontinuation occurred after an average of 4 semiannual visits. Thirty-seven percent (n = 211) of the women discontinuing stavudine did not report any ART at the time of stavudine discontinuation, whereas 56% (n = 322) of women discontinuing stavudine reported a HAART regimen at the time of stavudine discontinuation. Twenty-six percent (n = 150) of the women discontinuing stavudine reported zidovudine use at the time of stavudine discontinuation.
Annual Change in Body Weight and Circumference Measures: Univariate Analyses
Figure 1 illustrates the unadjusted estimated average annual change for weight and each circumference measurement concurrent with each primary exposure (HIV-uninfected and HIV-infected while using stavudine and HIV-infected after discontinuing stavudine). HIV-uninfected women demonstrated significant increases in weight (0.89 kg/y, 95% confidence interval [CI]: 0.37 to 1.40) and in each circumference (waist: 0.88 cm/y, 95% CI: 0.49 to 1.28; chest: 0.33 cm/y, 95% CI: 0.05 to 0.61; arm: 0.14 cm/y, 95% CI: 0.01 to 0.28; hip: 0.41 cm/y, 95% CI: 0.07 to 0.75; and thigh: 0.28 cm/y, 95% CI: 0.07 to 0.49). Continuing to report stavudine was associated with a significant decrease in average hip circumference (−0.86 cm/y, 95% CI: −1.49 to −0.22). Although decreases were also observed in weight and other measurements while continuing stavudine, these decreases were not statistically different from 0. Discontinuing stavudine was also associated with a significant decrease in hip circumference (−0.39 cm/y, 95% CI: −0.75 to −0.03) and nonsignificant changes in weight and other measurements.
Discontinuing stavudine was associated with smaller annual decreases compared with the changes seen while continuing stavudine for each of the 6 outcomes, although these differences were not statistically significant.
Annual Change in Body Weight and Circumference Measures: Multivariable Analyses
Figure 2 shows the adjusted estimated average annual change in weight and each circumference measurement for 5 groups: HIV-uninfected women; HIV-infected women who continued stavudine; and HIV-infected women who discontinued stavudine for ≤1 year, discontinued stavudine for >1 to 2.25 years, and discontinued stavudine for >2.25 years. Of the 574 women who discontinued stavudine during follow-up, 488, 337, and 218 contributed data at study visits occurring ≤1 year, between >1 and 2.25 years, and >2.25 years after stavudine discontinuation, respectively. After adjustment, HIV-uninfected women continued to demonstrate significant increases (P ≤ 0.01) in weight and in each circumference measure. Continuing to report stavudine use remained associated with a significant decrease in average hip circumference (−0.64 cm/y, 95% CI: −0.96 to −0.31) and nonsignificant decreases in weight and circumferences in the other 4 body sites.
The annual change in weight and body circumference measurements had a propensity to decrease less with increasing time after stavudine discontinuation. This effect was seen most prominently at the hip. The longest time without stavudine exposure (third tertile: >2.25 years) was associated with a significantly smaller average decrease in hip circumference (−0.06 cm/y) compared with the first (−0.64 cm/y) or middle (−0.46 cm/y) tertile. Furthermore, discontinuing stavudine for >2.25 years was associated with a significantly smaller average decrease in hip circumference than the annual decrease observed while on stavudine (−0.06 vs. −0.64 cm/y; P = 0.01). A similar pattern of results was observed at the thigh, where the average annual decrease (−0.005 cm/y) seen after >2.25 years off stavudine was close to 0 (P = 0.96) and significantly smaller than the annual changes seen among the first (−0.27 cm/y) and middle (−0.24 cm/y) tertiles.
We found that irrespective of whether stavudine was discontinued or continued, HIV-infected women continued to show decreases in weight and circumference measures of the waist, chest, hip, and thigh, whereas HIV-uninfected women demonstrated increases in these anthropomorphic outcomes. In addition, the annual declines in circumference measurements for HIV-infected women who discontinued stavudine tended to be less than those observed in women who continued stavudine. Only the average annual change in hip circumference >2.25 years after stavudine discontinuation was significantly different from the annual change while continuing stavudine, however. Furthermore, the average annual change in circumferences at the hip and thigh observed >2.25 years after stavudine discontinuation was statistically significantly different than the changes observed ≥1 year and between >1 and 2.25 years after discontinuation.
Compared with the changes seen while continuing stavudine, further decreases in the thigh circumference seemed to occur after discontinuing stavudine for ≤1 year, which could suggest a potential residual effect of stavudine even after discontinuation. It could be postulated that the further decreases in thigh circumference within the first year after stavudine discontinuation may be the result of a stavudine-induced mitochondrial toxicity, particularly in lower limb fat, which may persist for some time even when exposure to the drug is discontinued. Similar findings were also observed in a clinical trial of HIV-infected men.18 Leg fat measured by DXA was found to decrease to less than baseline values after 24 weeks off stavudine, but then to increase to greater than baseline values at 48 weeks off stavudine.18
It is also noteworthy that the annual change in arm circumference for HIV-infected women who discontinued stavudine was similar to that of HIV-uninfected women. Furthermore, the annual change in arm circumference for HIV-infected women seen while stavudine was used or after it was discontinued was not statistically different from 0, similar to our findings in the central sites of the waist and chest. These findings are different from those found in men, wherein arm fat seemed to increase significantly after discontinuation of stavudine compared with that in those who continued stavudine;5,6,8,18 increases in abdominal fat, however, were not uniformly observed in these studies. Findings from our study of HIV-infected women suggest that fat in the lower limb may be particularly affected by use of stavudine and that fat in the upper torso and waist may be least affected, which is contrary to early studies reporting that central lipohypertrophy predominated in HIV-infected women on ART.19-21
A primary limitation of our study is that we did not use direct measures of fat such as DXA, computed tomography (CT), or magnetic resonance imaging (MRI). Anthropometry is susceptible to nontrivial measurement error. To minimize variability in anthropometric measurements in our cohort, all clinicians performing anthropometry in the WIHS were trained and certified centrally, with recertification performed every 2 years. Another limitation is that at the baseline visit, fat loss was already evident in our HIV-infected women, who had significantly lower weight and circumference measures, particularly in the thigh and hip, compared with HIV-uninfected women. We also assume that the difference in the rate of change in anthropometry between those who continue or discontinue stavudine is attributable to changes in stavudine status per se. Therefore, we may not have seen more recovery of fat because the antiretroviral drug that replaced stavudine or a change in another medication in the regimen may have contributed to fat loss. These differences might explain why we found that the changes in HIV-infected women, for the most part, were significantly smaller than those observed in HIV-uninfected women. Furthermore, as with any observational study, it remains possible that our results are attributable to unmeasured factors related to stavudine discontinuation and changes in body circumference.
In summary, our findings suggest that HIV-infected women continue to lose fat, regardless of whether they continue or discontinue stavudine, compared with HIV-uninfected women, who continue to gain fat. The largest differences after stavudine discontinuation were observed in the hip, but the amount of recovery is modest relative to the increases observed in HIV-uninfected women. Therefore, in addition to discontinuing stavudine to reduce the rate of fat loss, other therapeutic strategies warrant further study in HIV-infected women.
Data in this study were collected by the WIHS Collaborative Study Group with centers (Principal Investigators) at the New York City/Bronx Consortium (Kathryn Anastos); in Brooklyn, NY (Howard Minkoff); at the Washington, DC Metropolitan Consortium (Mary Young); at the The Connie Wofsy Study Consortium of Northern California (Ruth Greenblatt); at the Los Angeles County/Southern California Consortium (Alexandra Levine); at the Chicago Consortium (Mardge Cohen); and at the Data Coordinating Center (Stephen J. Gange). Participating institutions approved this study and consent forms provided to study participants.
1. Mulligan K, Anastos K, Justman J, et al. Fat distribution in HIV
-infected women in the United States: DEXA substudy in the Women's Interagency HIV
Study. J Acquir Immune Defic Syndr
2. Tien P, Bacchetti P, Cofrancesco J, et al. Factors associated with regional adipose tissue in HIV
+ women. Presented at: 13th Conference on Retroviruses and Opportunistic Infections, 2006; Denver.
3. Tien PC, Cole SR, Williams CM, et al. Incidence of lipoatrophy
and lipohypertrophy in the Women's Interagency HIV
Study. J Acquir Immune Defic Syndr
4. Santos CP, Felipe YX, Braga PE, et al. Self-perception of body changes in persons living with HIV
/AIDS: prevalence and associated factors. AIDS
. 2005;19(Suppl 4):S14-S21.
5. Carr A, Workman C, Smith DE, et al. Abacavir substitution for nucleoside analogs in patients with HIV lipoatrophy
: a randomized trial. JAMA
6. John M, McKinnon EJ, James IR, et al. Randomized, controlled, 48-week study of switching stavudine
and/or protease inhibitors to Combivir/abacavir to prevent or reverse lipoatrophy
-infected patients. J Acquir Immune Defic Syndr
7. Martin A, Smith DE, Carr A, et al. Reversibility of lipoatrophy
-infected patients 2 years after switching from a thymidine analogue to abacavir: the MITOX Extension Study. AIDS
8. McComsey GA, Ward DJ, Hessenthaler SM, et al. Improvement in lipoatrophy
associated with highly active antiretroviral therapy
in human immunodeficiency virus-infected patients switched from stavudine
to abacavir or zidovudine: the results of the TARHEEL study. Clin Infect Dis
9. Moyle G, Sabin C, Cartledge J, et al. A 48-week, randomized, open-label comparative study of tenofovir DF vs abacavir as substitutes for a thymidine analog in persons with lipoatrophy
and sustained virologic suppression on HAART. Presented at: 12th Conference on Retroviruses and Opportunistic Infections, 2005; Boston.
10. Murphy R, Zhang J, Hafner R, et al. Peripheral and visceral fat changes following a treatment switch to a nonthymidine analogue or nucleoside-sparing regimen in patients with peripheral lipoatrophy
: 48-week final results of ACTG A5110, a prospective, randomized multicenter clinical trial. Presented at: 13th Conference on Retroviruses and Opportunistic Infections, 2006; Denver.
11. Gandhi M, Ameli N, Bacchetti P, et al. Eligibility criteria for HIV
clinical trials and generalizability of results: the gap between published reports and study protocols. AIDS
12. Bacon MC, von Wyl V, Alden C, et al. The Women's Interagency HIV
Study: an observational cohort brings clinical sciences to the bench. Clin Diagn Lab Immunol
13. Barkan SE, Melnick SL, Preston-Martin S, et al. The Women's Interagency HIV
Study. WIHS Collaborative Study Group. Epidemiology
14. Cole SR, Hernan MA, Robins JM, et al. Effect of highly active antiretroviral therapy
on time to acquired immunodeficiency syndrome or death using marginal structural models. Am J Epidemiol
15. Panel on Clinic Practices for Treatment of HIV
Infection. US Department of Health and Human Services and Henry J. Kaiser Family Foundation. Guidelines for the Use of Antiretroviral Agents in HIV-1 Infected Adults and Adolescents
. Bethesda, MD: National Institute of Health; 1998.
16. National Center for Health Statistics. The Third National Health and Examination Survey Reference Manuals and Reports
. Hyattsville, MD: National Center for Health Statistics, 1996.
17. Diggle P, Heagerty P, Liang K, et al. Analysis of Longitudinal Data
. 2nd ed. Oxford, UK: Oxford University Press; 2002.
18. Moyle GJ, Baldwin C, Langroudi B, et al. A 48-week, randomized, open-label comparison of three abacavir-based substitution approaches in the management of dyslipidemia and peripheral lipoatrophy
. J Acquir Immune Defic Syndr
19. Dong KL, Bausserman LL, Flynn MM, et al. Changes in body habitus and serum lipid abnormalities in HIV
-positive women on highly active antiretroviral therapy
(HAART). J Acquir Immune Defic Syndr
20. Galli M, Veglia F, Angarano G, et al. Gender differences in antiretroviral drug-related adipose tissue alterations. J Acquir Immune Defic Syndr
21. Gervasoni C, Ridolfo AL, Trifiro G, et al. Redistribution of body fat in HIV
-infected women undergoing combined antiretroviral therapy