Cardiopulmonary dysfunction in the form of decreased oxygen uptake at submaximal and peak exercise levels and a diminished workload capacity have been demonstrated in symptomatic and asymptomatic individuals infected with the human immunodeficiency virus (HIV-1) (18,27,32). Experimental studies with asymptomatic HIV seropositive patients indicate that aerobic exercise may yield a health benefit by direct improvement of cardiopulmonary and immune functioning (2,22). Adherence to exercise and other cognitive-behavioral interventions has also been associated with long-term HIV survival (17). Moreover, in a randomized controlled study of HIV infected and noninfected individuals who were blind to disease status, a moderate-intensity aerobic exercise program was found to buffer both psychological and immunological sequelae of a potent psychosocial stressor (21). Moderate intensity aerobic exercise has also been shown to significantly increase CO4 cell counts among asymptomatic HIV infected individuals (24). These studies suggest that not only may moderate aerobic exercise relieve current distress, it may also offer prophylactic benefits against immunological and psychological decline in response to future life-events that HIV-infected persons inevitably face. Additionally, results suggest that long-term benefits may only accrue from regular participation.
However, little is known regarding the effect of aerobic exercise and exercise adherence rate among symptomatic HIV-infected individuals. Several published studies assessing the effect of aerobic exercise training among HIV-symptomatic individuals reported that aerobic exercise improved cardiopulmonary function and quality of life (24,25,28,32,34). For example, a cross-sectional study of exercising and nonexercising persons infected with HIV who had progressed to either the symptomatic or acquired immune deficiency syndrome (AIDS) stages and who were receiving testosterone therapy found that exercisers significantly increased lean body mass and decreased depression and overall distress relative to nonexercisers (34). A recent randomized controlled 6-wk experimental study of HIV infected persons with CD4 cells between 100 and 500 revealed nonsignificant CD4 cell changes but significant improvements in aerobic fitness and quality of life in both moderate- and high-intensity exercisers as compared with nonexercising controls (32). However, although high-intensity exercise conditions produced the largest fitness effect, only moderate-intensity exercise produced significant immunological improvement in Candida albican antigen skin testing and a trend toward reduced plasma HIV RNA replication (32).
Although the data appear to support the efficacy of aerobic exercise to improve aerobic fitness particularly for asymptomatic HIV infected persons, the effect of exercise on symptomatic persons' CD4 cells is less clear. For example, one study has shown that aerobic exercise training normalizes CD4 cells, and it may buffer further decline particularly among psychologically distressed persons with HIV (22,24). In addition, benefit of exercise or absence of an iatrogenic effect may only extend to compliant exercisers. However, other studies (27,29,33) have not shown an increase in CD4 cells in this population. It has been suggested that the normalization of CD4 cells is mediated by decreases in psychological distress accompanying exercise (22). For example, although moderate exercise, in contrast to intense exercise, is generally considered to pose no immunological threat, it may be the case that even moderate aerobic exercise exerts an immunosuppressive effect among noncompliant individuals who are already manifesting symptoms of immunological compromise. That is, rather than achieving a physiological adaptation to exercise, symptomatic persons with HIV-1 who exercise irregularly may fail to physiologically adapt, and hence exercise may serve as a stressor.
The absence of a control group, sampling irregularities, or a lack of exercise relapse data in prior studies present further difficulties in the evaluation of aerobic exercise benefits for symptomatic individuals. For example, Rigsby et al.'s (28) and MacArthur et al.'s (25) samples primarily included participants who met current Center for Disease Control (CDC) (3) diagnostic criteria for AIDS based on CD4 lymphocyte count (< 200 mm3), whereas other studies have included heterogenous patient samples with respect to stage of HIV disease (24,25,32). Additionally, with few exceptions exercise dropout rates have typically not been reported (25,28,32). When exercise relapse rates are provided, the reasons for noncompliance, such as illness, were not typically stated. Lacking a uniform sample and adequate relapse information, it is difficult to assess the effects of exercise on patients at different disease stages or to determine whether illness occurring during the exercise protocol is associated with relapse rate. Sampling and exercise noncompliance issues are especially of concern considering that in one study relatively few patients (less 25%) completed the entire 24-wk program and less than half of the patients were compliant through 12 wk (25) and that the most recent well-controlled study to date reported a 23% dropout rate over 6 wk (32). Although this latter study concluded that dropout was not due to illness, the study was quite brief, and the patient sample was heterogenous with respect to disease status on the basis of CD4 cell count.
The purposes of the present study were to assess the effects of a 12-wk aerobic exercise program on aerobic fitness, CD4 cell count, and physician-assessed health status rating among symptomatic pre-AIDS HIV infected individuals and to assess the degree to which ill health was associated with exercise compliance. We hypothesized that individuals adhering to the exercise training program would improve aerobic fitness (O2peak) and physician-assessed health rating to a greater degree than noncompliant exercisers and a nonexercising control group. Frequency of illness was also hypothesized to be inversely associated with exercise compliance.
Participants. Participants were symptomatic HIV-1 seropositive men and women between the ages of 18 and 49 residing in South Florida who had been recruited for study participation from community centers, health clinics, and by newspaper advertisement between June 1993 and July 1995. Participants had been diagnosed with HIV-1 at least 6 months before study entry, and they showed clinical or laboratory signs consistent with mildly progressed HIV-1 infection but not AIDS (3). Of the 96 persons who inquired about the study, 53 did not meet eligibility requirements. Forty-three people meeting all entry criteria completed baseline assessments and were randomly assigned to either an exercise or a control condition, and 28 completed 3-month follow-up assessments. Of the 15 people not returning for 3-month assessments, six were originally assigned to the exercise group and nine were assigned to the control condition. Nonreturning exercise-assigned participants reported geographic relocation and time limitations due to starting school or a new job as reasons for not completing follow-up. Among nonreturning control participants, one was hospitalized, four could not be contacted or had moved, and four expressed general dissatisfaction with control group assignment.
Of the final sample completing both time point assessments, 5 women and 13 men were in the intervention condition, and 5 women and 5 men were in the control group. We compared study dropouts with participants who remained in the study on the psychological and physiological parameters measured at baseline to determine whether systematic bias existed. Dropouts were not significantly different from participants completing assessments at both time points with respect to age, body composition, race, fitness level, or CD4 cell count at baseline (all P > 0.30).
The final sample was comprised of 46.4% African American, 35.7% Caucasian, and 17.4% Hispanic participants with a mean age of 36.75 yr (SD = 6.27). Participants' body mass index (BMI) averaged 26.1 (SD = 6.45), and 32.1% of the sample currently smoked. Participants averaged 1.43 HIV-related physician-assessed symptoms at study entry. Symptoms included fevers, fatigue, lymphadenopathy, night sweats, persistent diarrhea, thrush, upper respiratory infections, shingles, bruises, muscle and joint pain, and rashes. There were no significant baseline differences between the intervention and control groups with respect to total number of symptoms or demographic variables (all P > 0.32).
Screening assessments. In compliance with university internal review board requirements, all potential subjects signed an informed consent and underwent a comprehensive screening procedure including a physical examination, blood draw, psychiatric evaluation, and a graded exercise test (GXT) before entry into the study. Participant screening was specifically designed to control for factors that might influence primary assessments and to ensure homogeneity of the sample with respect to HIV disease progression.
Inclusion criteria. Participants were required to be between 18 and 49 yr of age and to have had at least one non-AIDS related symptom, as determined by physical examination or laboratory signs of mildly progressed HIV infection (3). A 3-month stabilization period was instituted before study entry for participants discontinuing or beginning antiretroviral therapy, and no participants were receiving protease inhibitors. Similarly, participants currently engaged in psychotherapy or support groups were asked not to change their involvement during the study. The ability to participate in aerobic exercise and an average or below average fitness level, as determined by peak oxygen consumption (O2peak) during a stationary cycling GXT, was required of all participants (1,23).
Exclusion criteria. Subjects were excluded from participation if they: 1) were asymptomatic or had an AIDS-defining illness (3); 2) had a CD4 cell count < 200 mm3 or > 800 mm3; 3) reported drug or ethanol abuse within 6 months or had a positive toxicology screen; 4) met diagnostic criteria for a current major psychiatric disorder or were cognitively impaired as determined by an HIV specific Structured Clinical Interview for the DSM-III-R (31) and by Mini Mental Status Exam (10), respectively; 5) reported current use of psychotropic drugs or any medication that would limit functional capacity; 6) reported participation in regular aerobic or resistance-exercise training; 7) had any ACSM (1) contraindication for exercise testing or training or experienced any exercise test abnormality; or 8) had a chronic immune system depression related physical condition other than HIV such as cancer, chronic fatigue syndrome, hepatitis, or autoimmune disease. Additionally, participants who had been bereaved of a significant other within the previous 6 months and those scoring in the moderate to severe depression or anxiety ranges on the Hamilton Rating Scales for Depression and Anxiety (12,13), respectively, were also excluded from the study (15,16).
Health assessment. Physical examinations were conducted by a Florida-licensed physician with extensive experience in HIV symptomatology assessment, and blood samples for laboratory assays were drawn between 08:00 and 12:00 a.m. to control for potential diurnal variation. Physicians used a standardized physical assessment form to record symptom frequency and severity based upon CDC (3) category B criteria at both assessment points. Physician-assessed health status change from baseline to 3-month follow-up was then coded to reflect 1) deteriorated, 2) no change, or 3) improved. Coding was based on symptom frequency and severity as determined by physical exam and laboratory findings.
Laboratory assay. The presence of HIV-1 antibody was determined with the Abbott Laboratory (Abbott Park, IL) ELISA procedure. To confirm seropositivity status, samples that were positive by ELISA were repeated and subjected to an immunoblot using the methods described by the CDC. Measurement of T-helper-inducer (CD4+CD3+) cell count was chosen for its known sensitivity to aerobic exercise training (14,21,22,35) and for its clinical relevance to HIV-1 staging and disease progression (5). Only cells positive for both CD4+ and CD3+ markers, as determined by a whole blood two-color direct immunofluorescence procedure (9), were counted. Briefly, a commercially available monoclonal antibody conjugated with phycoerythrin was used to identify CD4+ cells. Aliquots of blood were incubated with the monoclonal antibody or isotopic control for 15 min at 4°C. Erythrocytes were lysed with an ammonium chloride-potassium bicarbonate-EDTA solution. After washing, the absolute count of positive cells was determined by flow cytometric techniques using Prism diagnostic soft-ware (Coulter; Hialeah, FL).
Graded exercise test. A symptom-limited GXT for determination of cardiopulmonary variables was performed on each subject within 1 wk before and 1 wk after the aerobic exercise training program. The GXT was performed using a SensorMedics (Yorba Linda, CA) Ergoline 800 electronically braked cycle ergometer, a SensorMedics 2900c metabolic cart, and a SensorMedics Marquette MAX-1 Stress ECG. As per ACSM guidelines (1), participants received a verbal overview of the GXT procedure and were fitted with a nose clip and a Hans Rudolph (Kansas City, MO) nonrebreathable mouthpiece to prevent the escape of expired air during the GXT. Briefly, the protocol was initiated at 30 W with a 30-W increase every 2 min thereafter, and participants were maintained at 60 rpm by digital display and metronome pacing. Criteria for GXT termination included two of the following: a plateau in maximal oxygen consumption, a respiratory ratio greater than 1.14, attainment of age predicted maximal heart rate, or voluntary discontinuation by the participant despite urging from the staff (1,24). Information generated from the GXT included: peak heart rate, peak oxygen consumption (O2peak), peak tidal volume (MTV), peak oxygen pulse (O2pulse), peak minute ventilation (E), and an indicator of leg power (Watts at peak exercise). Although a number of cardiopulmonary variables and a leg power indicator were obtained from the GXT, O2peak is considered to be the best overall indicator of aerobic fitness (23,30) and thus served as our primary marker of aerobic fitness.
Exercise training. Based upon prior work with HIV positive asymptomatic individuals (21-23), an interval cycling exercise program was developed for each participant assigned to the exercise group. Exercisers were required to come to our exercise laboratory three times a week for 3 months. Each 45-min exercise session was divided into alternating 3-min task periods, defined as 70-80% of maximal heart rate as determined by the GXT, and 2-min off-task recovery periods for a total of nine task and nine recovery periods per session.
Consistency of exercise sessions was ensured by the use of a Vantage XL wireless heart rate monitor (Polar, Port Washington, NY), which stored exercise session data and provided direct feedback to the participant and to laboratory assistants. Laboratory assistants individually monitored each participant to ensure that resistance on the cycle ergometer was adjusted to maintain heart rate in the target zone during task periods and to allow for adequate recovery during off-task periods. Vital signs were taken periodically throughout training sessions, and each participant was led through 10 min of stretching before and after each session.
Laboratory assistants were available for sessions from 6:30 a.m. to 8:00 p.m. Monday through Friday. To ensure that staff availability did not present a reason for exercise noncompliance, participants designated exercise session time, and participants were accommodated when they requested a schedule change. No participant indicated lack of exercise session availability as a reason for noncompliance.
Exercise compliance and reasons for relapse. Participants assigned to exercise training were queried regarding reasons for missed sessions. Missed training days were then coded to reflect either a planned absence (e.g., vacation or business trip), a sick day (e.g., fever or upset stomach), or an unplanned absence (e.g., working late or lack of motivation today). Missed sessions were coded as unplanned when participants did not report an illness or plan an absence. Exercise noncompliance was defined as 50% or more missed training sessions of the total number of sessions.
Procedures. After telephone and psychiatric screening, participants completed a physical exam, and a blood draw to confirm HIV-1 antibody status and to determine CD4 cell count eligibility. Participants successfully completing screening were then asked to return within the week to complete the GXT, which upon satisfactory completion was followed by a morning resting baseline blood draw 2-5 days later. Participants were then randomly assigned to either a waiting-list control or aerobic exercise training group. Participants assigned to the wait-list control condition received individual feedback regarding their GXT results and were asked to refrain from exercise for the duration of the study (3). Subjects assigned to the exercise condition followed the regimen as previously described. Health assessment and GXT procedures were repeated at the end of 3 months. Subjects were paid $75.00 for their participation at each time point.
Data analysis. As a first step, participants assigned to the exercise condition were further classified into compliant exercise and noncompliant exercise groups on the basis of exercise session attendance. Independent sample t-tests were then used to determine whether demographic or fitness variables at study entry differentiated compliant from noncompliant exercise-assigned participants. A chi-square analysis determined if number of smokers differed between groups. To determine whether illness during exercise training was related to noncompliance, a chi-square analysis compared the number of compliant and noncompliant participants who reported any bout of illness. Independent sample t-tests were also used to compare the mean number of sick days occurring during the training period for compliant and noncompliant groups.
Repeated measures ANOVA with group (compliant exercise, noncompliant exercise, and control) as a between-subjects factor and time point (baseline and 3-month) as the within-subjects factor determined the effect of the intervention on cardiopulmonary variables and CD4 cell count. With a repeated measures design, a significant time by group interaction is the effect of primary concern because it would indicate that any changes in cardiopulmonary function and CD4 cell count were differentially influenced by exercise condition. We confined our hypotheses testing to an examination of interaction effects that, when significant, were followed by paired t-tests to determine the magnitude of within-group changes. Because health status change represented an ordinal rank, a Kruskal-Wallis one-way ANOVA was calculated to determine the magnitude of between-group differences. An (0.05) alpha level was retained for all analyses. Because of small expected frequencies, a conservative Fisher's exact probability was calculated for all chi-square analyses (29).
Compliance and illness. A preliminary analysis indicated that 7 (38.9%) participants (4 men and 3 women) assigned to exercise training relapsed from exercise as defined by less than 50% attendance of possible sessions (noncompliant group). An examination of demographic and fitness variables at study entry indicated that compliant exercisers had lower BMI (mean = 22.75, SD = 2.95 and mean = 30.30, SD = 9.16 respectively) and smoked less than noncompliant exercisers (9.1% and 71.4%, respectively; P < 0.05). There were no significant differences between either exercise group and the control group with respect to smoking, demographic, and fitness variables at study entry (all P > 0.05).
Further analyses of self-reported illness as a reason for missing training sessions indicated that illness occurred with similar frequency in compliant (72.7%) and noncompliant (57.1%) exercise groups (Fisher's P = 0.63) and that all subjects who reported illness, with one exception, were medically cleared by the study physician to resume exercise. Mean illness days were also not significantly different between compliant (mean = 1.9, SD = 2.6) and noncompliant (mean = 6.6, SD = 3.3) participants, but unplanned misses were more likely among noncompliant (mean = 17.9, SD = 15.1) than compliant (mean = 4.5, SD = 3.9) participants (P = 0.012). Lastly, the number of exercise sessions missed due to illness was uncorrelated with entry CD4 cell count, r = −0.09.
Cardiopulmonary functioning, BMI, and leg power. With the exception of heart rate, repeated measures ANOVA revealed significant group by time-point interaction effects for all cardiopulmonary variables measured at peak exercise. Specifically, O2peak, O2pulse, MTV, and E were all significantly different (all P < 0.05). Follow-up paired t-tests used to clarify the nature of the interaction effects indicated that O2peak, O2pulse, MTV, and E increased significantly from baseline to 3-month follow-up among compliant exercisers (all P < 0.01, respectively; see Table 1). However, there were no significant changes in any cardiopulmonary variable (all P >0.05) for noncompliant and control subjects. Improvement in compliant exercisers varied from 12.2% to 13.8% in O2peak and O2pulse, respectively, and from 7.9% to 17% in MTV and E, respectively. A marginal group by time interaction effect was found for BMI (P < 0.06), but again, a significant increase in BMI was confined to the compliant exercise group (P < 0.05).
A significant group by time-point interaction effect was found for leg power, as indexed by Watts at peak exercise (P < 0.01). Leg power improved 24.7% in compliant exercisers (P < 0.01), but was not significantly changed in the other groups (P > 0.58).
CD4 cell count health status. A repeated measures ANOVA of CD4 cell count revealed a significant group by time interaction effect (P < 0.001; see Table 1). Follow-up paired t-tests indicated that compliant exercise participants significantly increased CD4 cell count by 13% (P < 0.02), whereas noncompliant exercisers significantly decreased CD4 cell count by 18% (P = 0.001) and control participants evidenced a decreasing trend of approximately 10% (P < 0.078).
Examination of physician-rated health status change revealed that only two participants were rated as improved (both compliant exercisers) at 3 months. Whereas deterioration in health appeared to vary among compliant (18.2%), noncompliant (28.6%), and control (55.6%) groups, the between group difference failed to reach statistical significance (P > 0.10).
The present study assessed the cardiopulmonary, immunological, and physician-assessed health status effects of a 12-week aerobic exercise program as well as the association between ill health and exercise relapse among individuals at an early symptomatic stage of HIV infection. Similar to prior studies indicating beneficial cardiopulmonary effects of aerobic exercise among HIV seropositive individuals (21,24,25,28,32), our findings support the hypothesis that participants adhering to a 12-wk aerobic exercise program could significantly improve O2peak to a greater degree than noncompliant exercisers and control participants. Moreover, despite a limited sample size, the effect size (Eta = 0.66) contrasting compliant exercisers with controls on O2peak was large as was the power to replicate (0.98). Compliant exercisers also significantly increased peak Watts, oxygen pulse, and ventilatory functioning, which were unaccompanied by adverse health status changes as determined by physician assessment. However, in contrast to hypotheses, the change in physician-rated health status from aerobic exercise intervention approached but did not achieve statistical significance, and the frequency of illness occurring during exercise training was unrelated to exercise noncompliance.
Our results further indicate that adherence to an aerobic exercise program may significantly increase CD4 lymphocyte count among symptomatic HIV-infected individuals. However, exercise noncompliance may be associated with faster CD4 cell decline. Noncompliant exercisers experienced a significant drop in CD4 cell count from baseline to 3-month follow-up. It is possible that inconsistent exercise attendance detracted from the noncompliant exercisers' ability to adapt to the physical strain of exercise, which may explain the drop in CD4 cell count. Among seronegative individuals, high-intensity exercise has been shown to produce acute immunosuppressive effects that diminish with subsequent training, and it is possible that even moderate-intensity exercise, of sporadic frequency, exerts a temporary immunosuppressive effect among those who are already immunocompromised (22,26). Our findings and those of Ironson et al. (17) suggest that exercise noncompliance may also be associated with a potential health risk for immunocompromised individuals. Specifically, Ironson et al. (17) reported that noncompliance with a cognitive behavioral stress management intervention was related to faster disease progression in a 2-yr follow-up study of HIV-infected gay men. Considering that immunological decrement may precede a physical decline and that poor treatment noncompliance has been associated with faster HIV disease progression (17), the CD4 cell decline among noncompliant exercisers may be of importance even though no significant group differences in physician-assessed health status were apparent at 3-month follow-up. Future immune reactivity studies with a wider immune profile may clarify whether moderate-intensity exercise is temporally immunosuppressive for HIV-1 symptomatic individuals and whether the CD4 cell decline found in noncompliant exercisers was an isolated finding or part of a larger prodromal phase accompanying a health status decline.
It is important to note that, although a bout of illness was a common occurrence in a symptomatic HIV population, neither number of illness days, health status, nor entry CD4 cell count differentiated compliant from noncompliant exercisers, whereas, similar to the general population, moderate obesity (BMI > 30) and smoking were determinants of exercise noncompliance (6,8). Noncompliant exercisers were also more likely to report unplanned events or a general lack of motivation as reasons for exercise session absence, and the exercise compliance rate (61%) among our HIV-1 symptomatic sample was comparable to the general population compliance rate (50%) (6) and to the compliance rate of a laboratory exercise study of equivalent length with an HIV population (25).
Lastly, although aerobic exercise programs have been shown to yield short-term cardiovascular and possible immune benefits for HIV-infected individuals (2,22-25,28,32), promoting the adoption and maintenance of exercise among the general population and among persons with chronic disease appears to be influenced more by behavioral, social, and psychological factors than by a lack of participant information regarding exercise prescription (8,33). A review of behavioral and psychological considerations identified substance abuse, fear of treatment side effects, practical inconvenience, and especially depression as primary factors related to treatment noncompliance among persons with HIV that also have relevance for exercise intervention (4,11). It has been suggested that targeting other risk reduction areas, especially smoking cessation and weight loss, in conjunction with an exercise program and the use of home-based intervention may be necessary to maintain exercise adherence among persons with chronic disease such as HIV infection (19-22). If exercise is to be used as an adjuvant treatment for individuals with symptomatic HIV infection, attention to factors influencing exercise adherence behavior is warranted.
1. American College of Sports Medicine. Guidelines for exercise test administration. In: Guidelines for Exercise Testing and Prescription,
4th Ed. Philadelphia: Lea and Febiger, 1991, pp. 55-92.
2. Calabrese, L. H., and A. Laperriere. Human immunodeficiency virus infection, exercise, and athletics. Sports Med.
3. Centers for Disease Control. Revised classification system for HIV infection and expanded surveillance case definition for AIDS
among adolescents and adults. Morbid. Mortal. Wk. Rep.
4. Chesney, M. A., and S. Folkman. Psychological impact of HIV disease and implications for intervention. Psychiatr. Clin. North Am.
5. Demartini, R. M., R. R. Turner, S. C., Formenti, et al. Peripheral blood mononuclear cell abnormalities and their relationship to clinical course in homosexual men with HIV infection. Clin. Immunol. Immunopathol.
6. Dishman, R. K. Overview. In: Exercise Adherence,
R. K. Dishman (Ed.). Champaign, IL: Human Kinetics, 1988, pp. 1-9.
7. Dishman, R. K. Increasing and maintaining exercise and physical activity. Behav. Ther.
8. Dubbert, P. M. Exercise applications and promotion in behavioral medicine. J. Consult. Clin. Psychol.
9. Fletcher, M. A., G. Baron, M. Ashman, M., Fischl, and N. Klimas. Use of whole blood methods in assessment of immune parameters in immunodeficiency states. Diagn. Clin. Immunol.
10. Folstein, M. F., S. E. Folstein, and P. R. McHugh. Mini Mental State: a practical method for grading cognitive state of patients for the clinician. J. Psychiatry Res.
11. Goodkin, K. Psychiatric aspects of HIV infection. Texas Med.
12. Hamilton, M. The assessment of anxiety states by rating. Br. J. Med. Psychol.
13. Hamilton, M. A rating scale for depression. J. Neurol. Neurosurg. Psychiatry
14. Hanson, P., and D. Flaherty. Immunological responses to training in conditioned runners. Clin. Sci.
15. Herbert, T. B., and S. Cohen, S. Stress and immunity in humans: a meta-analytic review. Psychosom. Med.
16. Herbert, T. B., and S. Cohen, S. Depression and immunity in humans: a meta-analytic review. Psychol. Bull.
17. Ironson, G., A. Friedman, N. Klimas, et al. Distress, denial, and low adherence to behavioral interventions predict faster disease progression in gay men infected with human immunodeficiency virus. Int. J. Behav. Med.
18. Johnson, J. E., G. T. Anders, H. M. Blanton, et al. Exercise dysfunction in patients seropositive for the human immunodeficiency virus. Am. Rev. Respirat. Dis.
141:618-622, 1990, 1990.
19. King, A. C., S. N. Blair, D. E. Bild, et al. Determinants of physical activity and interventions in adults. Med. Sci. Sports Exerc.
20. King, A. C., W. L. Haskell, D. R. Young, R. K. Oka, and M. L. Stefanick. Long term effects of varying intensities and formats of physical activity on participation rates, fitness, and lipoproteins in men and women aged 50 to 65 years. Circulation
21. Laperriere, A., M. H. Antoni, N. Schneiderman, et al. Exercise intervention attenuates emotional distress and natural killer cell decrements following notification of positive serologic status for HIV-1. Biofeedback Self-Regulat.
22. Laperriere, A., N. Klimas, M. A. Fletcher, et al. Change in CD4+ cell enumeration following aerobic exercise
in HIV disease: possible mechanisms and practical applications. Int. J. Sports Med.
23. Laperriere, A., M. H. Antoni, M. A. Fletcher, and N. Schneiderman. Exercise and health maintenance in HIV. In: Clinical Assessment and Treatment of HIV,
M. L. Galantino (Ed.). Thorofare, NJ: Slack, Inc., 1992, pp. 65-76.
24. Lox, C. L., E. McAuley, and R. S. Tucker. Exercise as an intervention for enhancing subjective well-being in an HIV-1 population. J. Sport Exerc. Psychol.
25. MacArthur, R. D., S. D. Levine, and T. J. Burk. Supervised exercise training improves cardiopulmonary fitness in HIV-infected persons. Med. Sci. Sports Exerc.
26. Nieman, D. C., A. R. Miller, D. A. Henson, et al. The effects of high- versus moderate-intensity exercise on natural killer cell cytotoxic activity. Med. Sci. Sports Exerc.
27. Posthoff, G., K. Wassermann, and M. Ostmann. Impairment of exercise capacity in various groups of HIV-infected patients. Respiration
28. Rigsby, L. W., R. K. Dishman, A. W. Jackson, G. S. Maclean, and P. B. Raven. Effects of exercise training on men seropositive for the human immunodeficiency virus-1. Med. Sci. Sports Exerc.
29. Rosenthal, R, and R. L. Rosnow. Essentials of Behavioral Research.
New York: McGraw-Hill Inc., 1984.
30. Saltin, B., and P. Astrand. Maximal oxygen uptake in athletes. J. Appl. Physiol.
31. Spitzer, R., J. B. W. Williams, M. Gibbon, and M. First. Structured Clinical Interview for DSM-III-R: nonpatient version for HIV studies (SCID-NP-HIV).
New York: Biometrics Research Department, New York Psychiatric Institute, 1988.
32. Stringer, W. W., M. Berezovskaya, W. A. O'Brien, C. K. Beck, and R. Casaburi. The effect of exercise training on aerobic fitness, immune indices, and quality of life in HIV+ patients. Med. Sci. Sports Exerc.
33. U. S. Department of Health and Human Services. Understanding and promoting physical activity. In: The Surgeon General's Report on Physical Activity and Health.
Washington, DC: Department of Health and Human Services, 1996, pp. 211-261.
34. Wagner, G., J. Rabkin, and R. Rabkin. Exercise as a mediator of psychological and nutritional effects of testosterone therapy in HIV+ men. Med. Sci. Sports Exerc.
35. Watson, R. R., S. Moriguchi, J. C. Jackson, L. Werner, J. H. Wilmore, and B. J. Freund. Modifications of cellular immune functions in humans by endurance exercise training during beta-adrenergic blockade with atenolol or propranolol. Med. Sci. Sports Exerc.