Smith, Barbara A.a; Neidig, Judith L.b,d; Nickel, Jennie T.e; Mitchell, Gladys L.c; Para, Michael F.b; Fass, Robert J.b
In the past few years there has been an improved understanding about the pathogenesis of HIV-1 infection, and as a result treatment has changed dramatically. New and more powerful drugs have been introduced and clinical trials demonstrate that combination antiretroviral therapies significantly delay the progression of HIV-1 and improve survival. In the US, the incidence of opportunistic infections is declining and mortality due to AIDS is decreasing .
Recently, however, optimism has been dampened somewhat by the report of persistence of HIV-1 transcription in peripheral blood mononuclear cells, despite suppression of plasma HIV-1 RNA copy number to undetectable levels for 20 months or more . The persistence of transcription, the continued presence of troubling symptoms such as fatigue, dyspnea, and undesirable weight change, and the recognition of new adverse effects associated with antiretroviral therapy (fat redistribution and metabolic change) require health care professionals to pursue more aggressive symptom management strategies.
The pharmacological approach to symptom management in HIV/AIDS patients is not without risk. Patients may not wish to add other medications to an already complex regime. Furthermore, some patients may wish to play an active role in promoting their own health and turn to alternative therapies .
Aerobic exercise training, which has been used in managing signs and symptoms of other chronic illnesses and is widely used in health promotion and rehabilitation programs to improve physical endurance, may be an effective alternative in the treatment of certain signs and symptoms associated with HIV-1 infection. Investigators have reported improved endpoints in healthy individuals and symptomatic improvements in individuals with coronary artery disease, cancer, and hypertension using aerobic exercise [4–10]. The first few studies of aerobic exercise training in HIV-1 individuals have demonstrated improved endpoints without decreasing CD4+ cell counts [11–13] or increasing HIV-1-RNA . Although these studies were limited by either the inability to directly measure total body oxygen consumption, small sample size or brief training protocols, the results suggest that aerobic exercise training could be an effective health promotion strategy for individuals with HIV-1 infection.
The overall aim of this study was to examine the effects of aerobic exercise training on parameters associated with fatigue, dyspnea, weight, and body composition. These three signs/symptoms were chosen because they were among those most often reported as troubling patients during visits to the clinic from which subjects were recruited . Our hypothesis was that aerobic exercise training, when used in this population, would decrease physiological fatigue (improve time on treadmill or physical endurance), reduce dyspnea, not precipitate wasting, and improve body composition without compromising an already fragile immune system. That is, we could exercise HIV-1-infected individuals without decreasing CD4+ cell count, CD4+ percentage, and without increasing plasma HIV-1 RNA copy number.
Materials and methods
The study was a randomized, wait-listed, controlled clinical trial that began enrolling patients in 1995 prior to the widespread use of highly active antiretroviral therapy (HAART). The last subject completed the exercise protocol in 1998. Study subjects were recruited from the Infectious Disease Clinic and an associated AIDS Clinical Trials Unit at a large mid-western, academic medical center as well as from local service and social groups serving HIV-1-infected individuals.
Sixty HIV-1-infected adults (200–499 × 106 CD4+ cells/l), 52 males and eight females, were randomized (1 : 1) to the experimental or the control condition. The sample size was based upon preliminary data from MacArthur et al.  and could detect a mean change of 2ml/kg per min. in maximum total body oxygen consumption (VO2 max) between the groups, with 87% power and a two-tailed alpha level set at 0.05. Subjects were on stable antiretroviral therapy; could not be taking anabolic steroids, growth hormone or appetite stimulants; and could not have had an AIDS-defining illness, fever, active wasting or weight < 85% of their ideal body weight. Female subjects could not be pregnant.
Institutional Review Board (IRB) approval was obtained, and subjects signed the approved consent form prior to enrolling. A research nurse screened subjects for eligibility using inclusion/exclusion criteria and a clinic physician performed a physical examination. At the Infectious Disease Clinic the subjects completed questionnaires, had body weight, skinfold thickness and circumferences measured, had blood drawn for hematology and chemistry laboratory measures, lymphocyte subsets and HIV-1 RNA, and were instructed by the research nurse to complete a 4-day food-use diary. At the Medical Center's Center for Wellness and Prevention, subjects completed baseline pulmonary function testing and the graded exercise test (GXT). Following completion of all baseline measures, subjects were randomly assigned to the experimental or control group and were continuously entered into the study in this manner until 60 subjects were enrolled.
Subjects assigned to the experimental group participated in an ongoing aerobic exercise-training program that was designed by study investigators, three times each week for 12 weeks. Subjects followed their individual exercise prescriptions under the close supervision of study staff during the regular operating hours of the Exercise Facility at the Medical Center. Subjects assigned to the control group were contacted every other week by telephone or during a clinic visit. At the end of the first 12 weeks of participation, subjects in the control group were enrolled in the exercise protocol and exercised to week 24.
In order to provoke continued adaptations in the cardiopulmonary system and skeletal muscle, subjects walked, jogged or ran on a treadmill or the track. The initial 20 min of each session were spent walking or jogging. Subjects could then use the stationary cycle, stair stepper or cross-country machine. This was done because the changes that occur in the skeletal muscle as the result of aerobic exercise training are specific to the muscle used. Thus, if subjects are tested while walking and running on a treadmill, a portion of the training should be dedicated to walking and running. The exercise was to be carried out 3 days each week for a minimum of 30 min at a workload that produced a heart rate corresponding to 60-80% of subject's VO2 max achieved on the GXT at baseline. Progressive increases in the exercise duration and/or intensity were made at intervals over the course of the training by adjusting the exercise workload to maintain the subjects’ heart rates within the initially prescribed ranges. Warm up and cool down exercises were performed in addition to the 30 min of aerobic exercise. Polar Vantage XL Heart Rate Monitors (Polar CIC Inc., Port Washington, New York, USA) were used during each exercise session so the subjects could be coached to keep their heart rates within the appropriate training range. The data were downloaded to substantiate that the subjects were exercising within their prescribed heart rate ranges.
Measurement of dependent variables
The dependent variables were measured at baseline and week 12 in all subjects and again at week 24 for the 15 control subjects who participated in the exercise program following completion of the control condition.
Time on treadmill
Time on treadmill was used as a marker of physical endurance (fatigue). Heart rate and electrocardiogram (ECG) were continuously monitored using a Medical Graphics Exercise ECG system (Medical Graphics Corp., St. Paul, Minnesota, USA) at rest and during each stage of the GXT protocol. When the subject grabbed the front rail of the treadmill indicating he/she could no longer continue, the treadmill was immediately returned to the starting speed and grade and the time on the treadmill was recorded automatically by the ECG monitoring system.
Rating of perceived exertion
The subjects were asked to rate exercise exertion on a scale of 6 to 19, the Borg rating of perceived exertion (RPE) scale, ranging from very, very light to very, very hard. RPE was used as a surrogate marker for dyspnea .
Forced expiratory volume at 1 s
Forced expiratory volume at 1 s (FEV1) was measured on each subject during a routine pulmonary function test to assess the presence of any functional lung changes that might contribute to the subjects’ experience of dyspnea.
The subject's height and weight, without shoes, were measured on a standard balance beam scale with a rigid vertical height rod and recorded.
Body mass index was calculated as weight (kg) divided by height2 (m). An estimation of total body fat was obtained by measuring the amount of subcutaneous fat using the thickness of specific skinfolds. Three central skinfolds (subscapula, suprariliac, vertical abdomen), and four peripheral skinfolds (triceps, biceps, thigh and medial calf) were measured three times to the nearest l.0 mm using Lafayette skinfold calipers. The means of the three measurements at each site were used to calculate a more precise estimate of the true value of the skinfold. Waist circumference at the umbilicus, a measure of central fat (subcutaneous and visceral) and maximum hip circumference were measured and recorded in mm. Waist : hip ratio (WHR) was calculated as waist circumference at umbilicus (mm) divided by maximum hip circumference (mm).
Maximal oxygen use/consumption
The subject's oxygen use was measured in a controlled laboratory setting by staff experienced with exercise testing in clinical populations using guidelines adapted from the American College of Sports Medicine Guidelines for Exercise Testing . The gas analyzers underwent a two-point gas calibration and a volume calibration using a 3-l calibration syringe. Barometric pressure, ambient temperature, and relative humidity were measured within 30 min of the GXT and used to standardize the measurement of exhaled volumes. The GXT was conducted using the OSU Hi-Fit treadmill protocol (The Ohio State University Center for Wellness and Prevention, Columbus, Ohio, USA) for testing adults with unknown exercise capacities. This testing protocol uses 2 min stages, begins at a speed of 2.5 mph, 0 grade, and the increase in the workload at each stage of the protocol is approximately midway between the standard Balke and Bruce treadmill protocols . Oxygen use (VO2), carbon dioxide production (VCO2) and respiratory exchange ratio (VCO2/VO2 in exhaled air) were continuously monitored. Heart rate, blood pressure and 12 lead ECGs were recorded at each workload.
All members of the GXT testing team, except the principal investigator, were blinded to the subject's group assignment and coached and encouraged the subject during the GXT. The principal investigator, when present to evaluate the consistency of the testing, did not coach or encourage the subject. The maximum oxygen use/consumption or VO2 max was considered to have been achieved when the subject: (a) demonstrated no further increase in VO2 with an increase in workload; (b) reached a respiratory exchange ratio (R) of > 1.09; (c) reached or exceeded 70% of their maximum voluntary ventilation; or (d) achieved a heart rate no less than 5% below their age-predicted heart rate (220 minus subject's age). The results were used to compare subjects on pre- and post-test measures of VO2 max; and to prescribe the appropriate exercise intensity for the experimental subjects.
CD4+ and CD8+ cell counts and percentages (lymphocyte subsets)
These parameters were measured following guidelines recommended by the Flow Cytometry Advisory Committee of the National Institute of Allergy and Infectious Diseases AIDS Clinical Trials Group Immunology Committee. To avoid the acute effects of exercise, samples were collected before or 48 h following the GXT.
The HIV-1 RNA polymerase chain reaction (Roche Amplicor HIV-1 Monitor Test, Roche Labs, Elizabeth, New Jersey, USA) became commercially available during the course of the study. Frozen plasma which had been stored at −70°C was available on 13 experimental and 24 control subjects at both baseline and week 12 and was assayed for HIV-1 RNA in a single batch.
Measurement of descriptive variables
Physical activity checklist and diary
General physical activity in which the subjects participated during the previous 2 weeks was measured at baseline and week 12 using an adaptation of an instrument originally designed by Paffenbarger et al. . Subjects in both groups were asked how many hours during the past 2 weeks they spent in light, moderate or vigorous activity. The primary purpose of this checklist and diary was to identify any disproportionate increase in moderate or vigorous non-training physical activity from baseline to week 12 that may have influenced study outcomes.
The Ohio State University HIV symptom checklist
For the purposes of this study, subjects in both groups were questioned at baseline about their symptoms using a 62-item self-report symptom checklist. The checklist was developed by study investigators using items from existing symptom checklists, HIV-1 symptoms reported in the literature, known drug side effects, and patient complaints identified in a preliminary study by Neidig et al. .
Four day food diary/nutrient intake
Nutrient intake was assessed at baseline and week 12. Subjects were asked to complete a 4-day food diary inclusive of a weekend day. Foods were coded and analyzed using N-Squared Nutritionist IV software (N-Squared Computing, Salem, Oregon, USA). Results are reported for total kilocalories consumed and percentage carbohydrates, protein and fat.
Data management and analysis
Investigators reviewed case record forms for missing data and aberrant values. Data were continuously entered and verified on a PC using Epi-Info data entry package (Centers for Disease Control and Prevention, Atlanta, Georgia, USA). Computerized accuracy checks were performed, errors were corrected, and data were transferred to the mainframe for analysis using SAS (SAS Institute Inc., Cary, North Carolina, USA).
Continuous baseline measures were compared between the exercise and control group using t-tests. If the data were not normally distributed the non-parametric Wilcoxon rank sum test was used. Categorical baseline measures were compared using a χ2 test. If the expected number of observations per cell was small, a Fisher's exact test was used.
Analysis of covariance (ANCOVA) was used to compare the change in each dependent variable measure at week 12 controlling for baseline values and other covariates. That is, covariates were retained in the model if the difference between the groups at baseline (P-value < 0.2) and if the association between the covariate and the dependent variable measure had a P-value < 0.2. Assumptions of the model were checked and met for all analyses .
To compare those in the exercise group who completed the initial 12 weeks of the study with those who did not complete it, a statistical approach similar to the approach used to compare the differences between the groups at baseline was used. That is, continuous baseline measures of those who completed the exercise training were compared with those who did not complete the exercise training using t-tests and categorical baseline measures were compared using a χ2 test. If the expected number of observations per cell was small, a Fisher's exact test was used.
All comparisons of changes in the dependent variables for control subjects who completed 12 weeks of exercise (n = 15) were performed using the Wilcoxon sign rank test, the non-parametric equivalent of a one-sample t test.
The two groups were similar at baseline on the variables of age, weight, BMI, (BMI in both groups exceeded 27), time since HIV-1 diagnosis, total number of symptoms reported, CD4+ cell count and percentage (Table 1). Seven subjects in each group were on protease inhibitor therapy at the beginning of the study. An equal number of women (n = 4) were randomized to each group. More African Americans and Hispanics were randomized to the experimental group (n = 10) in comparison with the control group (n = 3) (χ2 = 4.8; P = 0.03).
Forty-nine subjects (82%) completed the initial 12 weeks of the study: 19 experimental subjects and 30 control subjects. The 19 experimental subjects participated in a minimum of 28 (78%) of the 36 possible exercise sessions. Subjects’ reasons for not completing the study were related to employment, changes in schedule, transportation, family crisis and the time required for participation. No subject reported that they withdrew because of illness. As there was a disproportionate loss of subjects from the exercise group, we compared those subjects who completed the study with those who did not. The subjects who completed the study were not different from those who did not complete the study on age, weight, and number of symptoms reported. Nor were those subjects who withdrew early from the study sicker at baseline than those who completed the study based on CD4+ cell counts and percentage. That is, the CD4+ cell count and percentage were not different at baseline between those who completed and those subjects who did not complete the study. Additionally, we were able to informally follow many subjects who withdrew from the study, as they continued to receive care at the Infectious Disease Clinic.
Subjects who did not complete the study were more likely to have spent less time on the treadmill at baseline (7.2 ± 2.1 versus 9.2 ± 1.6 min; t = 2.82; P = 0.009) and to have a lower VO2 max (28.4 ± 9.4 versus 35.5 ± 6.1 ml/kg per min; t = 2.24; P = 0.04).
Table 2 reports the results of the ANCOVA analysis and includes only those subjects who have values for a variable at both baseline (pre) and week 12 (post). Potential confounding variables, identified by group comparisons as described under Data Management and Analysis included triceps skin fold, CD8 percentage, race and number of HIV medications at baseline. These variables were included in the ANCOVA analysis when appropriate. Table 3 reports the results of the analysis of the data from the fifteen control subjects who were enrolled in the 12-week supervised exercise program from week 12 to 24 following completion of their participation as control subjects.
Time on treadmill
The intervention (aerobic exercise training) in this study had a significant effect on time on treadmill (used as a marker of physical endurance/fatigue) after controlling for the initial baseline values and covariates. (Table 2) Those subjects in the experimental (exercise) group who completed the initial 12 weeks of the study were able to stay on the treadmill a mean of 1 min longer compared with the control subjects whose time on the treadmill remained virtually unchanged.
Subjects in the control group who exercised from week 12 to week 24 after completing their participation as a control subject (n = 15), increased their mean time on the treadmill from 8.5 to 9.2 min (P = 0.07) (Table 3).
RPE and FEV1
The intervention did not have a significant effect on the change in mean RPE during the GXT (P = 0.34) or on the change in mean FEV1, measured prior to the GXT (P = 0.32) (Table 2). Additionally the number of smokers (12 experimental, 17 control) was not different between the groups (χ2 = 0.601; P = 0.44), and there were no differences in RPE (t = −0.27; P = 0.79) or FEV1 (t = 0.18; P = 0.86) between smokers and non-smokers at baseline.
Weight and body composition
The intervention did have a significant effect on weight (P < 0.04), BMI (P < 0.04), triceps skinfold (P < 0.01), sum of the central skinfolds (P < 0.02), sum of the peripheral skinfolds (P < 0.02), circumference of the abdomen at the umbilicus (abdominal girth) (P < 0.02), and WHR (P < 0.003) after controlling for the initial baseline values and covariates (Table 2). That is, experimental subjects lost an average of 1.5 kg of body weight compared with control subjects who gained a mean of 0.5 kg. Neither the experimental nor the control subjects had a significant change from baseline to week 12 in total kilocalories consumed, F (1,26) = 2.18 (P = 0.15), or in the percentage of carbohydrates, F (1,26) = 0.04 (P = 0.84), or protein F (1,25) = 0.79 (P = 0.38) in their diet as reported in their 4-day food diary; however, the experimental subjects did significantly reduce the percentage of fat in their diet from 35% (baseline) to almost 30% (week 12) compared with control subjects who reduced the percentage fat in their diet from 36.5% (baseline) to 34.6% (week 12), F (1,23) = 4.51 (P = 0.045).
Subjects in the control group who exercised from week 12 to week 24 significantly decreased their body weight by 2.4 kg (Table 3). In addition, BMI was significantly reduced in this group from 28.3 to 27.5 as was the thickness of the triceps skinfold from 19.6 to 14.7 mm.
Maximum oxygen use/consumption
Those subjects in the experimental group who completed the initial 12 weeks increased their VO2 max by 2.6 ml/kg per min compared with the subjects in the control group who increased their VO2 max by 1.0 ml/kg per min; however, the trend towards improvement in the change in VO2 max did not reach statistical significance (P = 0.09) after controlling for initial baseline values and covariates (Table 2). The 15 control subjects who completed 12 weeks of exercise training following the control period also showed a trend toward improved VO2 max (+2.1 ml/kg per min; P = 0.11) from week 12 to week 24 (Table 3); however, that change in VO2 max also did not reach significance.
CD4+ cell count, CD4+ percentage and HIV-1 RNA copy number
The absolute number of CD4+ cells, percentage of CD4+ cells in blood, and the number of HIV-1 RNA copies in plasma were measured at baseline and week 12, in order to make certain that subjects could exercise without suppressing CD4+ cell counts or increasing plasma HIV-1 RNA. CD4+ cell counts, CD4+ percentage, or HIV-l RNA copy numbers did not change significantly during the study for the experimental or control groups (Table 2).
Twenty experimental and 18 control subjects were taking two or fewer antiretroviral medications at baseline and the number of subjects who had started protease inhibitor therapy, which became available during the course of this study, was the same (n = 7) in each group.
In the present study we hypothesized that we could safely exercise HIV-1-infected individuals and have a favorable impact on fatigue, dyspnea, weight and body composition.
Although the attrition rate for this study compares favorably with the rates of other exercise studies using an HIV-1-infected sample [11–13,15], there was a disproportionate loss of subjects from the exercise group. The 11 subjects who did not complete the study were similar in many variables to those who completed the study with the exception of time on treadmill and VO2max, which were significantly lower at baseline. As those subjects who were likely to benefit the most from the training were lost to follow-up, the results of the study could be biased toward the null hypothesis.
Despite this, exercise subjects were able to remain on the treadmill a minute longer, which represents an 11% improvement in exercise capacity. This is similar to the improvement seen in individuals treated with growth hormone for HIV-associated wasting . Parker et al.  found that a similar improvement in exercise outcomes resulted in a 40% reduction in the difficulty associated with activities of daily living. If the individual expends less effort in carrying out the activities of daily living, they are likely to be less fatigued.
Furthermore, this increase in time on treadmill occurred despite a reported increase in physical activity from baseline to week 12 by the experimental and control groups of 2.9 and 2.3 h, over the previous 2 weeks. These data are not consistent with Poehlman's finding that elderly subjects randomized to an exercise group report a decline in non-training physical activity , but data from both studies point to the need to monitor non-study-related activity during an exercise intervention.
We were interested in determining whether exercise would improve dyspnea, as shortness of breath was one of the three most troubling symptoms reported by patients in our clinic . However, subjects in the current study had a normal RPE response to the GXT and a normal FEV1 at baseline. This suggests a low prevalence of dyspnea in our sample and may account for our inability to document the impact of exercise on RPE and FEV1.
Many subjects who enrolled in the study were above ideal body weight. In addition to reducing weight, subjects reduced BMI, subcutaneous fat, central fat and WHR and did so without reducing kilocalories consumed.
Fat redistribution associated with HAART is characterized by increased central fat and peripheral fat loss [23–28]. In our study, we were able to demonstrate a significant reduction in waist circumference, the most robust anthropometric predictor of visceral adipose tissue . However, because our study included only 14 individuals on HAART more work needs to be done. In particular, it would be important to assess whether central fat can be preferentially reduced, which occurs in healthy individuals , without exaggerating the loss of peripheral fat.
VO2max is considered to be the best measure of cardiovascular capacity and cardiac output is the most important factor contributing to VO2max, assuming no pulmonary limitations. Skeletal muscle biochemical factors (oxidative enzyme activity and mitochondrial volume) are better predictors of endurance (time on treadmill) .
An explanation for the improvement in time on treadmill and not VO2max in this study may be the increased physical activity reported by the control group or to the withdrawal of those subjects who were likely to change the most. Both of these factors would have biased results toward the null hypothesis. An equally plausible explanation may be that there is a separation in muscle oxidative capacity and VO2max in this population. This has been shown in older women who exercised  and in muscle injury studies using magnetic resonance spectroscopy (MRS) . A study assessing VO2max, cardiac output and muscle biochemical factors may be particularly important in HIV-1-infected individuals and is supported by the findings of this study and a study by Mannix who reported a decrease in skeletal muscle oxidative capacity when AIDS patients were compared to healthy controls during an exercise task .
Aerobic exercise studies conducted earlier in the HIV-1 epidemic reported a significant increase in CD4+ cell counts [11,12]. The present intervention did not have a significant effect on CD4+ cell counts or HIV-1 RNA copies in either group, which is consistent with the findings by Stringer et al. . One explanation for the difference between the earlier studies and more recent studies may be related to the number and type of antiretrovirals the subjects were taking. The studies of LaPerriere et al.  and Rigsby et al.  both pre-dated the use of multiple drug combinations and HAART. Another explanation may be the timing of the blood sampling as exercise will acutely raise the number of circulating cells . In the current study the total number and type of antiretrovirals were similar between the groups and blood was drawn prior to or at least 48 h following the GXT.
Exercise has been shown to be safe in HIV-infected people because regular exercise has not been shown to activate an acute phase response and viral loads did not increase in response to acute exercise . The current study supports that exercise is safe as there was no significant change in the CD4+ cell counts or HIV-1 RNA copies in either group. In conclusion, supervised aerobic exercise training safely decreases fatigue, weight, BMI, subcutaneous fat and central fat in HIV-1-infected individuals.
The study investigators are forever indebted to Michael McDonald. We wish to acknowledge Dr. David Frid and the staff of the Ohio State University Center for Wellness and Prevention; research nurses, Jan Clark, Amy Fetzer, Jane Russell, Nancy Stark, and Kathy Watson; and exercise leader, Dylan Wessman, for their invaluable contributions to the study. Additionally, we wish to thank Amy Ferketich, Heather Houchard, and Nicole Sivak-Sears for their assistance with data analysis and Dr Marion Broome and Dr Michael Saag for their thorough reviews and critique of the many iterations of this manuscript. Finally, we want to remember Brenda Gaynor.
1. Palella FJ, Delaney KM, Mooreman AC. et al. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. N Engl J Med 1998, 338: 853 –860.
2. Furtado MR, Callaway DS, Phair JP. et al. Persistence of HIV-1 transcription in peripheral-blood mononuclear cells in patients receiving potent antiretroviral therapy. N Engl J Med 1999, 340: 1614 –1622.
3. Singh N, Squier C, Sivek C, Nguyen MH, Wagener M, Yu V. Determinants of nontraditional therapy use in patients with HIV infection. A prospective study. Arch Intern Med 1996, 156: 197 –201.
4. Casaburi RA, Patessio A, Ioli F, Zanaboni S, Donner CF, Wasserman K. Reductions in exercise in lactic acidosis and ventilation as a result of exercise training in patients with obstructive lung disease. Am Rev Respir Dis 1991, 143: 9 –18.
5. Dimeo F, Rumberger BG, Kaul J. Aerobic exercise as therapy for cancer fatigue. Med Sci Sports Exerc 1998, 30: 475 –478.
6. Fish AF, Smith BA, Frid DJ, Christmas SK, Post D, Montalto NJ. Step treadmill exercise training and blood pressure reduction in women with mild hypertension. Prog Cardiovasc Nurs 1997, 12: 28 –35.
7. Goldberg AP, Hagberg J, Delmez JA. et al. The metabolic and psychological effects of exercise training in hemodialysis patients. Am J Clin Nutr 1980, 33: 1620 –1628.
8. Hurwitz A. The benefit of a home exercise regimen for ambulatory Parkinson's disease patients. J Neurosci Nurs 1989, 21: 180 –184.
9. MacVicar MG, Winningham ML, Nickel J. Effects of aerobic interval training on cancer patients’ functional capacity. Nurs Res 1989, 38: 348 –351.
10. Preusser BA, Winningham ML, Clanton TL. High - vs low-intensity inspiratory muscle interval training in patients with COPD. Chest 1994, 106: 110 –117.
11. LaPerriere A, Fletcher MA, Antoni MH, Klimas NG, Ironson G, Schneiderman N. Aerobic exercise training in an AIDS risk group. Int J Sports Med 1991, 12 (Suppl. 1): S53 –S57.
12. Rigsby LW, Dishman RK, Jackson AW, MaClean GS, Raven PB. Effects of exercise training on men seropositive for the human immunodeficiency virus-1. Med Sci Sports Exerc 1992, 24: 6 –12.
13. Stringer WW, Berezovskaya M, O'Brien WA, Beck CK, Casaburi R. The effect of exercise training on aerobic fitness, immune indices, and quality of life in HIV+ patients. Med Sci Sports Exerc 1998, 30: 11 –16.
14. Neidig JL, Nickel JT, Smith BA, Brashers D, Para M, Fass B. Self reported symptoms of HIV infection. XI International Conference on AIDS. Vancouver, July 1996 [abstract Tu.B.176].
15. MacArthur RD, Levine SD, Birk TJ. Supervised exercise training improves cardiopulmonary fitness in HIV-infected persons. Med Sci Sports Exerc 1993, 25: 684 –688.
16. Borg GA. Psychological basis of perceived exertion. Med Sci Sports Exerc 1982, 14: 377 –381.
17. American College of Sports Medicine. Guidelines for Exercise Testing and Prescription. 5th edn. Baltimore, MD: Williams & Wilkins; 1995.
18. Paffenbarger RS Jr, Wing AL, Hyde RT. Physical activity as an index of heart attack risk in college alumni. Am J Epidemiol 1978, 108: 161 –174.
19. Snedecor GW, Cochran WG. Statistical Methods, 7th edn. Ames, IA: The Iowa State University Press; 1980.
20. Schambelan M, Mulligan K, Grunfield C. et al. Recombinant human growth hormone in patients with HIV-associated wasting: A randomized, placebo-controlled trial. Ann Intern Med 1996, 125: 873 –882.
21. Parker ND, Hunter GR, Treuth MS. et al. Effects of strength training on cardiovascular responses during a submaximal walk and a weight-loaded walking test in older females. J Cardpulm Rehabil 1996, 16: 56 –62.
22. Poehlman ET, Arciero PJ, Goran MI. Endurance exercise in aging humans: effects on energy metabolism. Exerc Sport Sci Rev 1994, 22: 251 –284.
23. Carr AS, Samaras K, Burton S. et al. A syndrome of peripheral lipodystrophy, hyperlipidaemia, and insulin resistance in patients receiving HIV protease inhibitors. AIDS, 1998, 12: F51 –F58.
24. Gervasoni C, Ridolfo AL, Trifiro G. et al. Redistribution of body fat in HIV-infected women undergoing combined antiretroviral therapy. AIDS 1999, 13: 465 –471.
25. Miller KD, Jones E, Yanovski JA, Shanker R, Feurstein I, Fallon J. Visceral abdominal-fat accumulation associated with use of indinavir. Lancet 1998, 351: 871 –875.
26. Shaw AJ, McLean KA, Evans B. Disorders of fat distribution in HIV infection. Int J STD & AIDS 1998, 9: 595 –599.
27. Shikuma CM, Waslien C, McKeague J. et al. Fasting hyperinsulinemia and increased waist-to-hip ratios in non-wasting individuals with AIDS. AIDS 1999, 13: 1359 –1365.
28. Walli R, Herfort O, Michl GM. 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.
29. Ross R, Fortier L, Hudson R. Separate associations between visceral and subcutaneous adipose tissue distribution, insulin and glucose levels in obese women. Diabetes Care 1996, 19: 1404 –1410.
30. Despres JP, Pouliot MC, Moorjani S. et al. Loss of abdominal fat and metabolic response to exercise training in obese women. Am J Physiol 1991, 26: E159 –E167.
31. Brooks GA, Fahey TD, White TP. Exercise Physiology: Human Energetics and its Applications. 2nd edn. Mountain View, California: Mayfield Publishing Company.
32. Newcomer BR, Larson DE, Bamman MM, Wetzstein CJ, Hunter GR. Muscle injury's effects on energy metabolism in a trained individual. Med Sci Sports Exerc 1999, 31: S71. S71.
33. Ullum H, Palmo J, Halkjer-Kristensen J. et al. The effect of acute exercise on lymphocyte subsets, natural killer cells, proliferative responses, and cytokines in HIV-seropositive persons. J Acquir Immune Defic Syndr Hum Retrovirol 1994, 7: 1122 –1133.
34. Roubenoff R, Skolnik PR, Shevitz A. et al. Effect of a single bout of acute exercise on plasma human immunodeficiency virus RNA levels. J Appl Physiol 1999, 86: 1197 –1201.
© 2001 Lippincott Williams & Wilkins, Inc.