|
HIV, ARTs, Aging & Body Changes
|
|
|
Longitudinal increases in waist circumference are associated with HIV-serostatus, independent of antiretroviral therapy
AIDS:Volume 21(13)20 August 2007p 1731-1738
Brown, Todd Ta; Chu, Haitaoa; Wang, Zhaojiea; Palella, Frank Jb; Kingsley, Lawrencec; Witt, Mallory Dd; Dobs, Adrian Sa
From the aJohns Hopkins University, Baltimore, Maryland
bNorthwestern University, Chicago, Illinois
cUniversity of Pittsburgh, Pittsburgh, Pennsylvania
dDavid Geffen School of Medicine at UCLA and the Los Angeles Biomedical Research Institute at Harbor-UCLA, Torrance, California, USA.
Authors: "In conclusion, we found that the increases in waist and hip circumferences among HIV-infected men over 5 years of follow-up and presumably attributable to aging were counterbalanced by opposing decreases in limb and trunk circumference measurements that were attributable to cumulative NRTI exposure. In addition, HIV-infected men showed more rapid increases in waist size compared to HIV-seronegative controls over the study interval, after adjustment for the effects of antiretrovirals. This latter observation may represent a 'return to health' phenomenon and requires further investigation.....Most of the individual PIs evaluated were not associated with statistically significant anthropometric changes... Each year of NRTI exposure was associated with a statistically significant 0.11 kg/m2 decrease in BMI.....cumulative NNRTI exposure was associated with a trend towards increasing BMI for each year (0.12 kg/m2).... Older age was associated higher baseline waist circumference (2.3 cm per 10-year increase, P < 0.001), but lower arm (-0.64 cm/10-year increase, P < 0.001) and thigh (-1.34 cm/10-year increase, P < 0.001) circumferences.... lower nadir CD4 cell counts were associated with lower BMI and smaller hip and thigh circumference measurements, and tended to be associated with smaller waist circumferences."
Abstract
Background: The relative contributions of the different classes of antiretroviral therapy (ART), HIV infection per se, and aging to body shape changes in HIV-infected patients have not been clearly defined in longitudinal studies.
Methods: Since September 1999, men enrolled in the Multicenter AIDS Cohort Study have undergone measurements of body mass index (BMI) and body circumferences at each semi-annual visit. The effect of HIV-serostatus and cumulative exposure to the three major ART classes on changes in anthropomorphic measurements occurring between 1999 and 2004 among HIV-infected and HIV-uninfected men were determined using linear mixed effects regression models.
Results: At baseline, average BMI and circumference measurements were greater in HIV-uninfected men (n = 392) than HIV-infected men (n = 661) (BMI, 27.3 versus 25.3 kg/m2; waist, 96.4 versus 90.2 cm; hip 101.3 versus 95 cm, thigh 54.1 versus 50.8 cm; arm 33.3 versus 31.7 cm, P < 0.001 for each comparison). Cumulative nucleoside reverse transcriptase inhibitor (NRTI) exposure, but not protease inhibitor or non-nucleoside reverse transcriptase inhibitor exposure, was associated with statistically significant changes in BMI (-0.11 ± 0.04 kg/m2 per year) and in circumferences of waist (-0.27 ± 0.07 cm/year), hip (-0.24 ± 0.05 cm/year), and thigh (-0.16 ± 0.03 cm/year) over the 5 years of follow-up. Independent of ART exposure, HIV-infected men had a more rapid increase in waist circumference over the study interval than did the HIV-uninfected men (difference 0.33 ± 0.15 cm/year, P = 0.02).
Conclusion: Cumulative NRTI therapy was associated with longitudinal decreases in body circumference measurements, whereas HIV-serostatus was associated with increases in waist circumference independent of ART.
BACKGROUND: In the era of highly active antiretroviral therapy (HAART), body habitus changes are common among HIV-infected patients, including fat accumulation in the visceral abdomen, neck and breast, and fat wasting in the extremities and face. Antiretroviral medications are important contributors to these changes. Nucleoside analogue reverse transcriptase inhibitors (NRTIs), particularly stavudine, have been especially implicated in peripheral fat wasting [1] with certain protease inhibitors (PIs) possibly having a synergistic effect [2]. Although the mechanisms underlying the pathogenesis of fat accumulation require further clarification, the use of PIs has been implicated [3].
With HAART initiation, HIV-infected patients often demonstrate increases in weight, trunk and limb fat over the first 16-24 weeks of therapy [2,4,5]. Subsequently, whereas amounts of trunk fat generally stabilize, limb fat frequently decreases progressively over time, which is most closely correlated with stavudine exposure [2,4,5]. Mallon, for example, reported a 13.6% decrease per year in limb fat after the initial 24 weeks of stavudine-containing HAART over 96 weeks among 40 HIV-infected men initiating therapy [4]. The stability of these body composition changes after prolonged HAART is unknown. In addition, the extent to which body composition changes occur as a result of aging or HIV infection per se is unknown.
To address these issues, longitudinal observation of body composition among HIV-infected and HIV-uninfected persons is necessary. In a cohort study of 815 women followed for 30 months in the Women's Interagency HIV Study (WIHS), HAART-experienced HIV-infected women were twice as likely to report peripheral lipoatrophy compared to HIV-seronegative controls, whereas incident lipohypertrophy was equally as common in HIV-infected and HIV-uninfected women [6]. In the Multicenter AIDS Cohort Study (MACS), we recently reported comparable increases in waist circumference among HIV-infected men receiving HAART and HIV-uninfected participants after adjustment for baseline characteristics. The rate of change for hip circumference was, however, less in HIV-infected men in comparison with HIV-uninfected controls (0.18 versus 0.49 cm/year) [7].
The aim of the current analysis was to use anthropometric measurements assessed at each MACS semi-annual visit among both HIV-infected and HIV-uninfected men to determine the relative contributions of the three major classes of antiretroviral therapy, aging, and HIV serostatus to observed changes in body composition over time.
Discussion
Our study had two major findings. First, in this group of highly antiretroviral-experienced males, cumulative exposure to NRTIs was associated with decreases in BMI and body circumference measurements over 5 years of follow-up, whereas cumulative exposure to either PIs or NNRTIs was not associated with changes in anthropometric measurements at any site. Second, we observed increases in BMI, waist circumference, and hip circumference among both HIV-infected and uninfected participants over the study interval, which likely reflect changes in body composition related to aging. Among HIV-infected men, however, waist circumferences increased more rapidly than in HIV-uninfected men, even after adjustment for the effects of antiretroviral exposure.
Our findings suggesting an association between NRTI use and decreases in body circumference measurements are consistent with other studies in the literature. Since early in the HAART era, NRTI therapy, particularly with thymidine analogs, has been associated with the development of measurable lipoatrophy [1,11]. For example, in a trial of 96 antiretroviral naive patients randomized to receive didanosine and stavudine with a protease inhibitor, after an initial increase in hip circumference, there was an average decline 0.18 cm/month over 32 months of follow-up [12]. In our study, HIV-infected men had already been exposed to NRTIs for a mean of 4.7 years prior to the baseline visit in 1999. Despite this long-term exposure, we found that each year of NRTI treatment, particularly treatment with stavudine, didanosine, and zidovudine, was associated with significant decreases in waist, hip, and thigh circumferences over the 5-year study interval, indicating that the anthropomorphic effects of NRTI therapy are progressive with prolonged exposure.
In contrast, we did not observe any significant anthropometric measurement changes related to either cumulative PI or NNRTI exposure. Although many antiretroviral-associated body composition changes have been observed since the introduction of PIs [13], their precise pathogenic role in adipocyte function among HIV-infected patients has been elusive. While PI exposure has been related to altered adipocyte differentiation through their effects on transcription factors such as SREBP-1 [14] and some clinical studies suggest that concomitant protease inhibitor therapy may worsen lipoatrophy when given in combination with thymidine analogs [1,2], other evidence suggests that the role of PI exposure in the lipoatrophy pathogenesis may be limited [15,16]. PIs have also been implicated in the pathogenesis of lipohypertrophy observed in some HIV-infected patients receiving HAART. We found no evidence that PI exposure was associated with increases in abdominal girth in our cohort.
Our second major finding dealt with comparative changes in body circumference measurements among HIV-infected versus HIV-uninfected men, independent of the effects of antiretroviral medication exposure. Unlike most previous studies of longitudinal changes in body composition in HIV-infected persons, our study included an HIV-uninfected control group, which helped us distinguish body shape changes attributable to HIV or its treatment from those attributable to aging. At the baseline visit in 1999, HIV-infected men in the MACS had smaller body circumference measurements at all sites in comparison with the HIV-seronegative men, as previously noted [17]. Although the etiology of these differences at baseline has not been specifically investigated in our cohort, the effects of chronic illness and possibly undernutrition in the HIV-infected patients prior to the availability of effective HIV treatment may play a role. Over the study interval, however, we observed significant increases in waist and hip circumferences in both HIV-infected and HIV-uninfected men consistent with the effects of aging, after adjusting for the effects of antiretrovirals.
To determine whether this adjusted increase in circumference measurements differed by HIV-serostatus, we included an interaction term in the statistical model between year since the baseline visit and HIV serostatus. We found that the average increase in waist circumference was more rapid among HIV-infected versus HIV-uninfected participants after adjustment for antiretroviral exposure. This is consistent with the hypothesis that a return to premorbid body composition occurs among HAART-treated HIV-infected men as proposed by Kotler [18].
Our study had several limitations. First, we used anthropometrics to measure changes in body composition. This technique can be subject to measurement error and cannot distinguish between changes in the various body compartments, such as subcutaneous and visceral compartments in the abdomen, as well as fat and muscle changes in the extremities. Prospective studies using MRI or CT imaging that include both HIV-infected and uninfected participants will be needed to clarify these issues. Finally, the MACS is a predominately white, male cohort, potentially limiting its generalizability.
In conclusion, we found that the increases in waist and hip circumferences among HIV-infected men over 5 years of follow-up and presumably attributable to aging were counterbalanced by opposing decreases in limb and trunk circumference measurements that were attributable to cumulative NRTI exposure. In addition, HIV-infected men showed more rapid increases in waist size compared to HIV-seronegative controls over the study interval, after adjustment for the effects of antiretrovirals. This latter observation may represent a 'return to health' phenomenon and requires further investigation.
Methods
Study population
The Multicenter AIDS Cohort Study (MACS) is an ongoing prospective study of the natural and treated history of HIV infection among homosexual/bisexual men in the United States [8]. Between 1983 and 1991, 5622 men were enrolled at four study sites (Baltimore, Chicago, Pittsburgh and Los Angeles) and participate in semi-annual study visits, which consist of a physical examination, laboratory testing, and detailed interview on the use of antiretroviral therapies in the preceding 6 months.
Beginning in April 1999 (31st MACS visit), circumference measurements of the waist, hip, thigh, and mid-arm were performed at each semi-annual visit. We restricted these analyses to MACS participants who participated in the 31st semi-annual visit and reported the information regarding their antiretroviral therapy use over the previous 6 months. Of the 5622 enrolled in MACS by 1991, 1857 HIV-uninfected men were administratively censored in 1996 and 1750 men had died by 1 April 1999, leaving 2015 individuals. Of these, 1064 (53%) participated in the 31st semi-annual visit. After we excluded 11 men who did not report information on antiretroviral therapy use, the final study population comprised 1053 MACS participants. These analyses were based on up to 5 years of follow-up from the 31st semi-annual visit to the 41st semi-annual visit, which occurred between April and September 2004.
Anthropomorphic measures
A wall-mounted stadiometer was used to measure height. Each participant was weighed while wearing minimal clothing or an examination gown. All circumferences were measured with the participant standing in a relaxed position using the protocol established in the Third National Health and Nutrition Examination Survey (NHANES III) [9]. To minimize variability in measurement technique, each examiner at the four sites underwent the same videotaped instruction and personalized training with a recognized expert in anthropometric measurements (W.C. Chumlea, Wright State University, Dayton, Ohio, USA).
Assessment of antiretroviral therapy exposure
Cumulative exposure to each of the three major antiretroviral classes [protease inhibitors (PI), nucleoside reverse transcriptase inhibitors (NRTI), and nonnucleoside reverse transcriptase inhibitors (NNRTI)] was determined based on self-reported antiretroviral therapy (ART) use in the previous 6 months and quantified using time-varying continuous variables denoting years of use for each therapy class beginning 3 years before the baseline visit (visit 31). Our aim in using this 3- year cut-point, which corresponds approximately to the time at which PIs were introduced, was to confine NRTI exposure to the HAART era. Since NRTIs had been introduced several years earlier, we made this adjustment so that the duration of NRTI exposure since study entry would not become a 'de facto' marker of HIV disease duration as previously described [10]. Age, body mass index (BMI; weight in kilograms divided by height in meters squared) and nadir CD4 cell count were ascertained for all participants at the baseline visit.
Statistical analysis
To assess the effects of antiretroviral therapy on body shape changes over time, multiple linear mixed effects regression models were implemented using the SAS PROC MIXED procedure (SAS Institute, Cary, North Carolina, USA). The dependent variables were the anthropomorphic measures: BMI, circumferences of waist, hip, arm and thigh. The independent variables included the cumulative years of exposure to each of the three drug classes (i.e. PI, NRTI and NNRTI), as well as HIV serostatus, time in years since the index visit, and their interaction. For adjustment, age (centered at 45 years) and nadir CD4 cell count (centered at 265 cells/ƒÊl), both ascertained at the index visit, were used as covariates for all models. For HIV-uninfected men, we did not include the effect of nadir CD4 cell count or cumulative antiretroviral exposure in the models. Baseline BMI was included as a covariate for all anthropometric measures except for itself. These models were fitted to the data with random intercept reflecting individual differences at baseline (visit 31) and random slope of time reflecting individual rate of changes, unstructured covariance of random coefficients, and constant error variances. The restricted maximum likelihood method was used for the estimation of regression coefficients. These models provide estimates of the average linear trajectories over time, while accounting for correlation among repeated measurements from the same participants.
The person-visits of which any outcome or covariate were missing were excluded from the multiple mixed effects regression models. After this exclusion, 942 participants (who contributed 6599 person-visits), 936 participants (who contributed 6533 person-visits), 935 participants (who contributed 6525 person-visits), 937 participants (who contributed 6543 person-visits), and 936 participants (who contributed 6522 person-visits) were included in the models of BMI, circumferences of waist, hip, arm and thigh, respectively.
We then undertook an exploratory analysis to determine whether cumulative exposure to specific antiretroviral medications in each class was associated with changes in anthropometric measurements. We restricted this analysis to HIV-infected men and examined those specific antiretroviral medications used most often at the baseline visit. Each antiretroviral medication was tested in a separate model. In addition to baseline age, BMI, nadir CD4 cell count, and time from baseline, each model was adjusted simultaneously for cumulative exposure to the other two classes of antiretroviral medications.
Results
Baseline characteristics
The baseline characteristics of the HIV-uninfected and HIV-infected groups are shown in Table 1. The HIV-uninfected group was older and had a higher mean BMI than the HIV-infected group. Mean body circumference measurements were significantly larger at baseline at all sites measured among the HIV-uninfected men versus the HIV-infected group. HIV-infected patients had mean cumulative drug exposures of 2.26 years for PIs, 4.78 years for NRTIs, and 0.51 years for NNRTIs at baseline. Cumulative baseline exposures to the most commonly used medications within each of these classes are also provided in Table 1.
![StudyIndex-1.gif](../images/081507/StudyIndex-1.gif)
Cumulative exposure to antiretroviral class and longitudinal changes in anthropometry
BMI was affected differently by cumulative exposure to each of the three major antiretroviral classes after adjustment for HIV serostatus, nadir CD4 cell count and age. Each year of NRTI exposure was associated with a statistically significant 0.11 kg/m2 decrease in BMI (95% CI: -0.18, -0.04 kg/m2, P = 0.001). In contrast, cumulative PI exposure was not associated with significant BMI changes, while cumulative NNRTI exposure was associated with a trend towards increasing BMI for each year (0.12 kg/m2; 95% CI, -0.02, 0.26, P = 0.09).
Figure 1 shows the average adjusted changes in waist, hip, arm, and thigh circumferences for each year of exposure to the three classes of antiretroviral medications. Each additional year of NRTI exposure was associated with a significant decrease in waist, hip, and thigh circumference measurements (waist -0.27 cm, P < 0.001; hip -0.24 cm, P < 0.001; thigh: -0.16 cm, P < 0.001). Each year of PI and each year of NNRTI exposure were associated with non-significant increases at all anatomical sites measured.
Fig. 1. Annual change in anthropomorphic measurements and 95% confidence intervals for each additional year of exposure to the three major classes of antiretroviral medications in HIV-infected men: PI (protease inhibitor), NRTI (nucleoside reverse transcriptase inhibitor), NNRTI (nonnucleoside reverse transcriptase inhibitor). (a) waist circumference, (b) hip circumference, (c) arm circumference, (d) thigh circumference.
![changeWaist-2.gif](../images/081507/changeWaist-2.gif)
Influence of other covariates on anthropometric measurements
Table 2 shows the baseline anthropometric measurements (intercept) and the effect of age, BMI (for circumference measurements), HIV-status, and nadir CD4 cell count (HIV-infected only) on the intercept values. Older age was associated higher baseline waist circumference (2.3 cm per 10-year increase, P < 0.001), but lower arm (-0.64 cm/10-year increase, P < 0.001) and thigh (-1.34 cm/10-year increase, P < 0.001) circumferences. BMI was associated with higher baseline waist, hip, arm, and thigh circumferences. HIV-serostatus was associated with significantly lower baseline values for all anthropometric measurements, except arm circumference. Among HIV-infected men, lower nadir CD4 cell counts were associated with lower BMI and smaller hip and thigh circumference measurements, and tended to be associated with smaller waist circumferences.
BMI measurements increased significantly in HIV-uninfected men (0.204 ± 0.04 kg/m2 per year, P < 0.001) as did waist (0.512 ± 0.08 cm/year, P < 0.001) and hip (0.512 ± 0.08 cm/year, P < 0.001) circumference measurements. We observed statistically non-significant decreases in mean arm and thigh circumferences (data not shown). In order to determine whether observed anthropometric measurement changes were different among HIV-infected versus uninfected men, we evaluated the interaction between HIV-status and number of years elapsed since baseline. Waist circumferences increased more rapidly in HIV-infected patients (0.33 ± 0.15 cm/year, P = 0.02) than HIV-uninfected controls (Fig. 2). In addition, there was a trend toward more rapid increases in thigh circumference among HIV-infected versus HIV-uninfected participants (0.13 ± 0.08 cm/year, P = 0.07). Further adjustment for duration of HAART therapy yielded similar results (data not shown).
Effect of cumulative exposure to individual antiretroviral medications on anthropometric outcomes
Table 3 shows the changes in each of the anthropometric measurements among HIV-infected men for each year of exposure to individual antiretroviral medications within the three major drug classes after simultaneous adjustment for cumulative exposure to the other two classes and other baseline factors. Most of the individual PIs evaluated were not associated with statistically significant anthropometric changes. Exceptions were the effect of nelfinavir on arm circumference (-0.15 ± 0.15 cm/year of exposure, P = 0.045) and saquinavir on arm circumference (+ 0.17 ± 0.14 cm/year of exposure, P = 0.019). Among the NRTIs, each year of exposure to zidovudine was associated with decreases in arm (-0.16 ± 0.09 cm, P < 0.001) and thigh (-0.12 ± 0.1 cm, P = 0.018) circumferences. Each year of stavudine exposure was associated with a significant decrease in BMI (-0.09 ± 0.09 kg/m2, P = 0.036), waist (-0.27 ± 0.19 cm, P = 0.004) and hip (-0.34 ± 0.14 cm, P < 0.001) circumferences. Didanosine exposure was associated with decreases in BMI (-0.21 ± 0.11 kg/m2 per year of exposure, P = <0.001) and hip circumference (-0.22 ± 0.19 cm/year of exposure, P = 0.024). Tenofovir use was associated with decreases in hip circumference (-0.35 ± 0.34 cm/year of exposure, P = 0.046), but increases in arm circumference (+1.24 ± 0.5 cm/year of exposure, P < 0.001). None of the NNRTIs were associated with significant changes in the anthropometric outcomes.
Table 3. Effects of cumulative drug years of individual drugs in each class on each anthropomorphic outcomea in HIV-infected men.
![Ampren-3.gif](../images/081507/Ampren-3.gif)
|
|
|
|
|
|
|