Drs. Hatoum and Walton noted the presence of some baseline differences between groups in our study. There certainly were some differences among the groups, as would be expected by chance in any small randomized clinical trial. The most potentially important difference was in highly active antiretroviral therapy (HAART) use. If anything, a rate of HAART use of 88% in subjects in the oxandrolone (OX) group might have biased that group toward more improvement (compared with 72% in nutrition alone [NA] subjects and 69% in progressive resistance training [PRT] subjects). Baseline values for most key variables (CD4 count, viral load, fat-free mass [FFM], thigh cross-sectional muscle area [CSMA], and physical functioning) were nearly identical for OX and PRT. We think it is unlikely that the other small and nonsignificant differences noted are clinically important (eg, body mass index [BMI] of 20.6 kg/m2 in the PRT group vs. 19.8 kg in the OX group). It is unclear how the cited article by van Nes et al,1 a study of non-HIV-infected Swiss elderly persons, relates to this work.
Drs. Hatoum and Walton expressed concern about missing visits and adherence to study protocols. Space constraints prevented us from fully describing the procedures we used to support such adherence. These included financial incentives, snacks during visits, transportation fare, weekly telephone calls from familiar personnel, reminder calls before visits, and pocket calendars. Study personnel were in constant contact with the participants. Still, some appointments (more often in the exercise arm) were missed because of personal responsibilities, family and personal problems, and health issues. If anything, this would have biased the results toward less effectiveness from exercise. Subject retention in the study was excellent (94%) and similar among the treatment arms, however, and we do not believe that differences were sufficient to bias between-group comparisons.
Drs. Hatoum and Walton commented that our primary study end point was functional change and that this was biased toward PRT because of a “learning effect.” We disagree. The primary end points were thigh muscle mass and self-reported physical functioning (a quality of life index). For a learning effect to have taken place, participants in the PRT arm would have had to receive the outcome assessment tool on more occasions or somehow have practiced responding to these items, which they did not. This is true of the muscle strength testing, which was not an end point but rather meant to demonstrate effective specific muscle training from PRT, which it did. It is well known that specific muscle strengthening does not necessarily lead to parallel changes in muscle mass, lean body mass, other strength and performance measures, or self-report; hence, the need for this study.
Drs. Hatoum and Walton correctly point out that OX and PRT effectively increased CSMA and FFM more than NA. Although potentially important, these findings were only true when tested within-arm. None of these changes was significantly greater than the change in the NA arm, our control group. Although interesting to look at, the differences between the OX and PRT arms were not the goal of this study because of sample size limitations, and they were neither significant nor large enough to emphasize. Among other nonprimary outcomes, and almost more notably, improvements in dietary intake were the smallest, and statistically so, in the OX group. As pointed out, the fall of health-related quality of life (HRQL, a primary outcome) in the OX arm was the most unexpected and striking finding.
Drs. Hatoum and Walton also question the methods we used to assess utilities. The study design included assessment of utilities using traditional time tradeoff methods. Unfortunately, patients were unable to understand these items, and we could not therefore use these standard utility measures. Instead, scores on HRQL obtained using the HIV costs, services and utilization study (HCSUS) instrument,2 were translated to a single health state classification or quality-adjusted life-year (QALY) measure for each client. The measure incorporates 6 dimensions of health, including physical functioning, role limitation, social functioning, bodily pain, mental health, and vitality. The cross walk involved applying methods used by Brazier et al3 to derive a preference-based single index from the Short Form-36 (SF-36). Because the individual items on the HCSUS and SF-36 are nearly identical, using this method seemed completely justifiable, although we acknowledge, as did Brazier, that the crosswalk requires minor assumptions for remaining differences. Regression models based on a visual analog scale and a standard gamble method were developed using these 6 domains to estimate the final QALY measure for this study. For our analysis, we chose the visual analog scale approach for estimating a single index QALY measure; using this method, results were consistent with findings from other study outcomes.
This study was designed, funded, and begun before the publication of Strawford et al4 suggested that combination therapy might be the most beneficial approach. Doubts remain about the applicability of those results, however, because of the supraphysiologic dosing of androgens used and because the findings of Grinspoon et al5 did not demonstrate such a benefit of combined treatment modalities. We do not consider exercise and OX mutually exclusive; they could certainly be used together. Nevertheless, patients, providers, and payers all need to understand the effectiveness, limitations, and costs of each approach. In developing countries, the cost of OX (which must be imported) would be substantially more than that of PRT (which would rely on local labor). In many industrialized countries, budget constraints are severe, most notably the state portion of the Medicaid program. Given that this may be the last opportunity to conduct such a study of fairly pure wasting, we believe we were obligated to draw some conclusions that might help providers, institutions, and third-party payers with their decision making.
Abby H. Shevitz, MD, PhD*†
Ira B. Wilson, MD†‡
Ann Y. McDermott, PhD§
Donna Spiegelman, PhD∥
Sarah C. Skinner, MA*
Kristina Antonsson, MD, MPH¶
Jennifer E. Layne, MS#
Aaron Beaston-Blaakman, PhD**
Donald S. Shepard, PhD**
Sherwood L. Gorbach, MD*
*Nutrition Infection Unit Department of Community Health Tufts University Boston, MA
†Department of Medicine Tufts-New England Medical Center Boston, MA, ‡Institute for Clinical Research and Health Policy Studies and the Department of Medicine Tufts-New England Medical Center Boston, MA, §Lipid Metabolism Laboratory Jean Mayer US Department (USDA) of Agriculture Human Nutrition Center on Aging at Tufts University Boston, MA, ∥Departments of Epidemiology and Biostatistics Harvard School of Public Health Boston, MA, ¶Department of Medicine Stamford Hospital Stamford, CT, #Nutrition, Sarcopenia, and Exercise, Physiology Laboratory Jean Mayer USDA Human Nutrition Center on Aging at Tufts University Boston, MA, **Schneider Institute for Health Policy Heller School, Brandeis University Waltham, MA
1. Van Nes MC, Herrmann FR, Gold G, et al. Does the mini nutritional assessment predict hospitalization outcomes in older people? Age Ageing
2. Hays R, Cunningham WE, Sherbourne CD, et al. Health-related quality of life in patients with human immunodeficiency virus infection in the United States: results from the HIV cost and services utilization study. Am J Med
3. Brazier J, Usherwood T, Harper R, et al. Deriving a preference-based single index from the UK SF-36 Health Survey. J Clin Epidemiol
4. Strawford A, Barbieri T, Van Loan M, et al. Resistance exercise and supraphysiologic androgen therapy in eugonadal men with HIV-related weight loss. JAMA
5. Grinspoon S, Corcoran C, Parlman K, et al. Effects of testosterone and progressive resistance training in eugonadal men with AIDS wasting. A randomized, controlled trial. Ann Intern Med