| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Department of Psychology (M.H.A., A.W.C., R.E.D., S.S., F.P., G.I., N.S.), Department of Psychiatry (M.H.A., G.I., N.S.), and Department of Medicine (M.A.F., N.K., N.S.), University of Miami, Coral Gables, FL.
Address correspondence and reprint requests to Michael H. Antoni, PhD, Department of Psychology, 5665 Ponce De Leon Blvd, Coral Gables, FL 33146. E-mail: Mantoni{at}miami.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Methods: HIV-positive men who have sex with men were randomized to either a 10-week CBSM + MAT intervention (n = 76) or a MAT-Only condition (n = 54). Data were collected at baseline immediately following the 10-week intervention period, at 9 months postrandomization, and at 15 months postrandomization.
Results: We found no differences in HIV viral load among the 130 men randomized. However, in the 101 men with detectable viral load at baseline, those randomized to CBSM + MAT (n = 61) displayed reductions of 0.56 log10 units in HIV viral load over a 15-month period after controlling for medication adherence. Men in the MAT-Only condition (n = 40) showed no change. Decreases in depressed mood during the intervention period explained the effect of CBSM + MAT on HIV viral load reduction over the 15 months.
Conclusions: A time-limited CBSM + MAT intervention that modulates depressed mood may enhance the effects of HAART on suppression of HIV viral load in HIV+ men with detectable plasma levels.
Key Words: depression • HIV/AIDS • HIV viral load • intervention
Abbreviations: AIDS = acquired immunodeficiency syndrome; ACTG = Adherence to Combination Therapy Guide; BDI = Beck Depression Inventory; CBSM = cognitive behavioral stress management; CET = coping effectiveness training; EDTA = ethylenediaminetetraacetic acid; HAART = highly active antiretroviral therapy; HIV = human immunodeficiency virus; MAT = medication adherence training; MEMS = medication event monitoring system; POMS = Profile of Mood States.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Many of the psychosocial and behavioral difficulties experienced by HIV-positive persons may be addressed with stress management and behavioral interventions that improve functioning across a variety of domains (13). In particular, group-based interventions appear to be well suited for HIV-positive populations as they provide opportunities for enhanced skill acquisition and received social support (14). We have previously reported that group-based cognitive behavioral stress management (CBSM) teaches stress reduction skills, decreases depressed mood, provides social support, and is associated with increases in indicators of immune system reconstitution in a cohort of mildly symptomatic men who have sex with men (15,16). Specifically, men reported reductions in distress and increases in perceived social support during the 10-week intervention period, which appeared to explain the effect of CBSM on a measure of antiviral immunity (17,18). Other trials have observed that HIV-positive men in group-based coping effectiveness training (CET) reported lower perceived stress and burnout, as well as higher coping self-efficacy following the 10-week intervention period, when compared with an HIV informational control group (19). Over the maintenance phase, where men in both conditions received six booster sessions, CET effects on enhanced positive morale were sustained through the 12-month follow-up. It is noteworthy, however, that these trials were conducted before the availability of HAART, and hence the findings may not generalize to contemporary cohorts of HIV-positive men who have sex with men.
In the HAART era, a variety of behavioral interventions for HIV-positive persons has also been developed specifically to support medication adherence. Although psychological adjustment has not uniformly been conceptualized as a mechanism of enhanced adherence outcomes, research has supported the efficacy of pharmacist-led individualized medication adherence training (MAT) interventions and those that employ cognitive-behavioral principles based on self-efficacy theory (20). We theorized that a modified form of CBSM may offer benefits in improving mood, health behaviors, and immune status in the era of HAART. As a result, the present trial was designed to test the added value associated with providing stress management training by comparing the combination of CBSM and MAT (i.e., CBSM + MAT) to MAT alone in a cohort of HAART-treated HIV-positive men who have sex with men. We hypothesized that the CBSM + MAT intervention would decrease HIV viral load more than MAT-Only over the 15-month investigation period. Intervention-related increases in adherence were hypothesized to be one plausible mechanism of reductions in HIV viral load. Another was intervention-related decreases in depressed mood, a known predictor of HIV disease progression, which was a major focus of this study. In order to examine intervention-related reductions in depressed mood as a potential mediator of intervention effects on disease progression, we tested the effects of CBSM + MAT on HIV viral load against a MAT-Only condition and after controlling for adherence at four time points over a 15-month observation period. By examining CBSM + MAT effects after controlling for adherence training and adherence levels, we were able to determine whether there was any remaining variance in viral load reductions which could be associated with changes in depressed mood.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
, 
globulin), (B) had a history of chemotherapy or whole-body radiation treatment for cancer that was not AIDS-related, or (C) had a history of chronic illness associated with permanent changes in the immune system. Temporary exclusion criteria included antibiotic use for an acute infection within the past 2 weeks, changes in the HAART regimen or acute bodily infection during the past month, hospitalization for surgery within the past 3 months, and intravenous drug use within the previous 6 months. Additional exclusionary criteria, determined at the initial appointment, included (A) cognitive impairment; (B) inability to read at a sixth-grade level; (C) current psychosis, drug or alcohol dependence, and panic disorder; and (D) active suicidality. During the initial visit, men between the ages of 18 and 65 who reported no changes in their HAART regimen during the past month completed an informed consent, provided morning peripheral venous blood samples, and received a physical examination from a physician. Participants were assessed using the HIV Dementia Scale (21) and selected modules of the Structured Clinical Interview for DSM-IV (22) to determine eligibility. Eligible participants were randomized into either a 10-week CBSM + MAT intervention or a MAT-Only condition. To facilitate within group analyses, we randomized approximately 40% more of the eligible participants to CBSM + MAT. Randomization procedures were conducted by a master's-level project manager and overseen by the principal investigator. Participants' identification numbers were drawn from a box for assignment to the study conditions. Participants completed self-report measures of mood and adherence, as well as provided morning peripheral venous blood samples at baseline (i.e., month 0), during the week following the intervention period (i.e., 3 months), at 9 months postrandomization, and at 15 months postrandomization. Research personnel were blind to participants' experimental condition. At baseline, participants were provided with a medication event monitoring system (MEMS) cap and continued to meet with a licensed clinical pharmacist at 3 months and 9 months postrandomization to download these in conjunction with MAT sessions held during assessment visits. At each of the four assessment visits, men were provided with a $50 incentive for their participation, as well as $10 compensation for travel expenses ($60 total).
CBSM Intervention
Participants attended 10 weekly 135-minute group sessions (90 minute stress management and 45 minute relaxation) and were instructed to practice relaxation exercises twice daily between sessions. Guided by a detailed training manual, postdoctoral fellows and advanced clinical-health psychology graduate students led groups of four to nine participants. Group facilitators were supervised weekly by a licensed psychologist or board-certified psychiatrist using audiotapes to monitor intervention fidelity. Sessions included a didactic component, as well as group discussion, with opportunities provided to apply newly learned techniques. Homework was assigned to provide opportunities for participants to practice techniques and increase self-efficacy.
This modified form of CBSM focused extensively on eliciting participant experiences with adherence and medication side effects. Throughout the 10 sessions, facilitators encouraged participants to examine potentially distorted cognitions and how these may influence adherence to HAART (as well as other relevant self-care behaviors). During cognitive restructuring exercises, participants were asked to examine medication-relevant thoughts both in session and through homework exercises. Adherence was also a key target during the skills training sessions. For example, participants were encouraged to break down the larger stressor of adherence into smaller, more manageable components and then match these to a productive coping response (i.e., emotion-focused versus problem-focused). Finally, participants were also taught a variety of relaxation exercises, including progressive muscle relaxation, autogenic training, meditation, and deep breathing. Relaxation exercises were presented as an important form of emotion-focused coping that may be particularly effective in managing HAART side effects (e.g., peripheral neuropathy). Although 10 CBSM sessions were conducted, men attended a median of six CBSM group sessions. Men randomized to CBSM + MAT received a $5 incentive for their participation and $10 compensation for travel expenses ($15 total) at each of the group sessions.
MAT Intervention
All participants received MAT from a licensed clinical pharmacist in a 1-hour session at baseline, as well as half-hour maintenance sessions at 3 months and 9 months postrandomization, respectively. MAT was delivered in conjunction with regularly scheduled assessment visits. All 130 participants attended at baseline, 98 (CBSM + MAT = 64; MAT-Only = 34) attended at 3 months, and 77 (CBSM + MAT = 51; MAT-Only = 26) attended at 9 months. The MAT aims to increase information about HIV and HAART, including how the medications work, why they must be taken on time and at the proper dose, how to recognize possible side effects, and explore medication-related concerns with one's health care provider (23). Information is provided in a context that attempts to build on intrinsic motivation for change by addressing potential strategies that will support adherence to HAART. Two weeks after the posttreatment assessment, MAT-Only participants were offered a 1-day educational seminar that consisted of a condensed version of the 10-week CBSM. This was done to ensure that all men had access to information regarding stress management and relaxation training, which may be helpful in the management of HIV disease. At the 1-day seminar, men randomized to MAT-Only received a $5 incentive for their participation and a $10 compensation for travel expenses ($15 total). In total, 38 men attended the 1-day seminar.
Recruitment and Retention
From 1998 to 2004, we approached 257 HIV-positive men who have sex with men via fliers at health care facilities, as well as through referrals from current participants, physicians, and HIV/AIDS service organizations (see Figure 1). After a preliminary telephone contact, 81 refused participation due to scheduling conflicts, travel requirements, and lack of interest. The 81 men who refused participation after an initial telephone contact did not significantly differ in age, education, employment status, and time since diagnosis from those who elected to participate (all p values > .10). Of the 176 participants who expressed further interest, 46 were excluded for the following reasons: positive urine screen for illicit substances (n = 12), presence of a comorbid medical condition known to affect immune function or no current prescription for antiretroviral medications (n = 9), active suicidal ideation (n = 6), panic disorder (n = 5), AIDS-related dementia (n = 4), less than a sixth-grade reading level (n = 3), antisocial personality disorder (n = 2), active psychosis (n = 2), men who were not gay/bisexual or participating in another behavioral intervention trial with HIV-positive men (n = 2), and refused randomization (n = 1). Thus, we began the trial with 130 men who were randomized to CBSM + MAT (n = 76) and MAT-Only (n = 54).
|
Throughout the present trial, we employed a number of commonly utilized strategies to enhance retention of participants (24). At consent, study staff emphasized the long-term follow-up period for the trial and highlighted the importance of participation for helping others with HIV/AIDS. We also provided the aforementioned incentives for participation, reimbursement for travel expenses, and small tokens of our appreciation (e.g., holiday cards) to participants. At baseline, each participant provided a weekly schedule, and we made every effort to accommodate group sessions, as well as study assessment visits to match availability. A detailed log of contacts for each participant was kept in a secure location, and participant appointments were tracked with a computer-generated calendar. However, due to the fact that data collection required morning peripheral venous blood samples drawn by medical staff at our Behavioral Medicine Research Center, we were unable to schedule assessment appointments at offsite locations (e.g., a participant's home).
For the present investigation, a participant was classified as a completing follow-up if he returned at either 9 or 15 months postrandomization (n = 78; 60%). Of those who returned during the follow-up period, 77 provided data at 9 months and 71 at 15 months. These retention rates are comparable to those reported in a trial of CET with HIV-positive men who have sex with men (i.e., 67%) where maintenance data were collected through a 12-month follow-up (19). Interestingly in that study it appears that all participants received six booster sessions throughout the maintenance phase. This enhanced contact with the study team across the conditions may have resulted in slightly higher retention rates when compared with the present trial where men received only one half-hour MAT session at 9 months postrandomization. For those men who did not complete follow-up, staff spoke with 7 who indicated they were no longer interested in participating. In addition, staff were unable to contact 45 men via telephone or mail (CBSM + MAT = 22; MAT-Only = 23) to schedule follow-up assessments.
Control Measures
Demographics
Age, ethnicity, education, income, relationship status, sexual orientation, and religion were assessed.
Health Status
Lowest historical CD3+ CD4+ cell counts, HIV symptoms, prescription of psychotropic medications, prescription of acyclovir, and time since HIV diagnosis were assessed during a physical examination with study physicians at baseline.
AIDS Diagnosis
As per the current Centers for Disease Control and Prevention criteria, study physicians made a diagnosis of AIDS with historical or current evidence of a CD3+ CD4+ count below 200 cells/mm3 or an AIDS-related condition (25).
Medication Adherence
The Adult AIDS Clinical Trial Group Adherence to Combination Therapy Guide (ACTG) was used to assess self-reported HAART adherence (26). Interviewers asked each participant to report the number of doses skipped for each day over the past 4 days. The percentage of pills missed was subtracted from 100 to calculate percent adherence.
MEMS caps contain a pressure-activated microprocessor which records the date, time, and duration of each opening. A licensed clinical pharmacist (collaborating with study physicians) selected the most potent antiretroviral in a participant's regimen to monitor using the MEMS. MEMS cap data were electronically downloaded (MEMS View version 2.61; Aprex Corporation). Percent adherence was calculated for three periods: (1) between baseline and 3 months; (2) 90 days following the 3-month assessment; and (3) from 90 days after the 3-month assessment through 9 months postrandomization. We also calculated percent adherence with the MEMS over the 4 days before each assessment (3 months and 9 months postrandomization) to examine data comparable to that obtained using the ACTG.
Depressed Mood and Anxiety
The Profile of Mood States (POMS), a 65-item scale, assesses six different mood states during the past week using a variety of adjectives that participants were asked to rate using a 5-point Likert-type scale: 0 = not at all to 4 = extremely. The POMS has internal consistencies close to or greater than 0.90, as well as high external validity (27). We selected the POMS to measure depressed mood and anxiety because these subscales are not confounded with HIV symptoms. Both the depressed mood and anxiety subscales displayed adequate internal consistency in this sample (Chronbach's = 0.94 and 0.78, respectively).
The Beck Depression Inventory (BDI), a 21-item scale, assesses somatic, affective, behavioral, and cognitive dimensions of dysphoria during the past 2 weeks (28). The BDI is reliable and has shown good concurrent validity with the Hamilton Rating Scale for Depression (0.73 to 0.80 for nonpsychiatric patients). The internal consistency of the BDI for the present investigation was adequate (Chronbach's = 0.90). In addition, we examined selected symptoms measured by the BDI as a means of cross-validating any observed effects of CBSM + MAT on depressed mood as measured by the POMS. Specifically, we selected BDI items assessing sadness, hopelessness, dissatisfaction, and guilt because these most closely reflected the adjectives used in the POMS-depressed mood subscale. When summed, these items displayed adequate internal consistency (Chronbach's = 0.72) and were examined as a modified BDI-Affective composite score (range = 0–12). This score was viewed as less likely than the total BDI score to be confounded by physical disease status because it was free of somatic symptoms.
HIV Disease Status
CD3+ CD4+ T-Lymphocyte Counts
Morning peripheral blood samples were collected from all subjects in ethylenediaminetetraacetic acid (EDTA) tubes (Vacutainer-EDTA, Becton-Dickinson, Rutherford, NJ). CD3+ CD4+ lymphocyte count was determined by whole blood four-color direct immunofluorescence using a Coulter XL and flow cytometer (29).
HIV Viral Load
HIV-1 viral load was determined on EDTA plasma using an in vitro reverse transcriptase polymerase chain reaction assay (AMPLICOR, Roche Laboratories, US #83088). This ultrasensitive assay has a lower limit of 50 copies/mL.
Statistical Analyses
Analyzing repeated measures in randomized clinical trials can present challenges due to the increased likelihood of having missing data at distal follow-up points. A major concern with using split-plot, repeated-measures ANOVA to analyze effects of interventions is that because it requires listwise deletion of cases, this methodology yields potentially biased parameter estimates due to attrition over time. The mixed-model methodology provides a valid test of the intent-to-treat hypothesis because it does not require listwise deletion and as a result is able to use all available data at each assessment time (30). This methodology can also account for unequal time intervals, makes less stringent assumptions regarding within-subject correlation patterns and enables one to utilize parameter estimates of change per 1-month increase in time to predict dependent variable values. Predicted values are a function of unstandardized parameter estimates based on the likelihood of obtaining the observed data. In preliminary, mixed-model analyses for this study, fixed effects were time (either 0, 3, 9, and 15 months or just 0 and 3 months), group (CBSM + MAT versus MAT), and the interaction of group x time. Where we observed CBSM + MAT effects on depressed mood or adherence in preliminary mixed-model analyses, these were to be examined as potential mediators of treatment effects on HIV viral load using the MPlus statistical package for structural equation modeling (31). We specified a latent growth curve model with the trajectory of change in HIV viral load over time (i.e., slope) and HIV viral load at 15 months (i.e., intercept) as two latent variables, both of which are required for model identification. The 15-month values for HIV viral load were specified as the intercept for this model by employing accepted modifications to the time structure (i.e., –15, –12, –6, 0). Thus, we could examine the mechanisms of change in HIV viral load and outcomes at 15 months. This flexible, innovative statistical methodology has been reviewed in depth elsewhere (32).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
2 and independent samples t-tests indicated that completers did not differ significantly from noncompleters in age, ethnicity, religion, education, income, sexual orientation, time since diagnosis, lowest historical CD3+ CD4+ T cell counts, and number of HIV-related symptoms at baseline (all p values >.10). The CBSM + MAT condition was implemented in 12 separate cohorts of participants. Cohort status (i.e., 1–12) was unrelated to changes in HIV viral load, measures of mood, and adherence across the 15-month investigation (all p values >.20).
Using structural equation modeling, we examined the association between the 2 measures of adherence (i.e., ACTG and MEMS) and HIV viral load separately. We specified a latent variable of MEMS using the three available indicators measured over the investigation period and examined its association with a latent variable of HIV viral load (which was composed of four indicators over the 15-month investigation period). We determined that this model was a good fit for the data (2 [13] = 15.289, p = .29; CFI = 0.99; RMSEA = 0.037). Results indicated that the latent variable of MEMS was not significantly related to the latent variable of HIV viral load (ß = –0.052, t = –1.10, p >.05). Next, we specified a latent variable for the ACTG (composed of four indicators over the 15-month period) and examined its association with the latent variable of HIV viral load. We determined that this model was an adequate fit for the data (2 [18] = 24.10, p = .15; CFI = 0.97; RMSEA = 0.051). In contrast to MEMS data, the latent variable of ACTG was significantly related to the latent variable of HIV viral load (ß = –1.558, t = –3.18, p < .01). Thus, self-reported adherence (i.e., ACTG) appeared to be the most valid method of assessment in this sample of HIV-positive men who have sex with men.1
HIV Viral Load and CD3+ CD4+ T-Cell Counts
After controlling for adherence, we examined intervention effects for HIV viral load and CD3+ CD4+ T-cell counts across the 15-month investigation period and found no differences between conditions (p > .10). Next, we conducted a secondary analysis by excluding 29 men with undetectable viral load at baseline. It is plausible that we were unable to detect potential intervention-related reductions in HIV viral load due to floor effects in those men who already presented with undetectable levels at baseline.
CBSM + MAT Effects in Men With a Detectable HIV Viral Load
A majority of the 101 participants with a detectable HIV viral load at baseline were non-Hispanic white (52%), but a sizeable minority were African American (21%) and Hispanic (20%). A majority of the sample (84%) indicated they were predominantly or exclusively homosexual in orientation. Most participants were single (65%) and reported being raised with Christian religious traditions (59%). Mean age was 41.6 (SD = 8.3) years, and the sample was largely well educated, with 75% of participants having completed at least some college. The majority of participants (70%) reported that they were not working at study entry. The modal income was $5,000 to $20,000 per year. The average time since HIV diagnosis was 7.8 (SD = 5.1) years, and participants reported an average of 6 (range = 0 to 12) HIV symptoms. The majority of participants (54%) were diagnosed with AIDS at baseline. The mean baseline CD3+ CD4+ T cell count was 422 (SD = 240) cells/mm3, and we observed a mean viral load of 17,047 (SD = 33,938) copies/ml at baseline. We observed no baseline differences between the CBSM + MAT and MAT-Only conditions in demographics, health status variables, AIDS diagnosis, adherence, depressed mood, or HIV viral load (all p values >.10).
Among this subsample of 101 men, completers did not differ significantly from noncompleters in ethnicity, religion, education, employment status, income, sexual orientation, time since diagnosis, psychotropic medications, AIDS category C status, number of HIV-related symptoms, lowest historical CD3+ CD4+ T-cell counts, CD3+ CD4+ T-cell counts, and HIV viral load (all p values >.10). Completers (M = 43.8, SD = 9.3 years) were significantly older than noncompleters (M = 38.4, SD = 8.3 years; p < .01). Consistent with previous findings (8), younger age was associated with decreased adherence at baseline (r = 0.30, p < .01) but was unrelated to HIV viral load, CD3+ CD4+ counts, or depressed mood (all p values >0.10).
HIV Viral Load and CD3+ CD4+ T-Cell Counts
For the 101 men with detectable plasma HIV levels at baseline, we observed a significant group x time interaction for HIV viral load across the investigation period after controlling for HIV medication adherence (ß = –0.056, t(131) = –2.10, p < .05). Men in CBSM displayed significant reductions in HIV viral load through the 15 months (ß = –0.037, t(131) = –2.40, p < 0.05), whereas controls displayed no change (ß = 0.02, t(131) = 0.87, p > 0.30).2 Men in CBSM + MAT displayed a 0.037 log10 reduction in HIV viral load per month, yielding a 0.56 log10 reduction in HIV viral load (i.e., more than a three-fold decrease) over the investigation period (see Figure 2). In fact, a 0.50 log10 decrease in HIV viral load has been identified as the minimum criterion for determining the short-term effectiveness of a HAART regimen (33). Descriptive statistics for observed, untransformed HIV viral load are summarized in Table 1. It is noteworthy that in contrast to those assigned to MAT-Only, men in CBSM + MAT achieved median viral loads that were below the detectable range of our assays at 15 months. Within the CBSM + MAT condition, attending more group sessions was associated with greater reductions in HIV viral load, ß = –0.004, t(100) = –2.48, p < .05. At the same time, however, we found no intervention-related changes in CD3+ CD4+ T-cell counts over the 15 months (p > .50).
|
|
Medication Adherence
Among men with detectable viral load at study entry, the experimental conditions did not differ significantly in participant-reported medication adherence throughout the 15-month investigation period, even after controlling for age (p > .50). Each group reported nonsignificant increases in adherence over time. At baseline, approximately 70% of men reported levels of adherence that met or exceeded 95% of those medication doses prescribed. No differences between the experimental conditions were observed throughout the investigation period for the number of participants reporting greater than or equal to 95% adherence (p values >.10). Observed values for measures of medication adherence across the 15-month investigation period are reported in Table 2.
|
Depressed Mood
For the POMS-Depression subscale, we observed a significant group x time interaction over the 10-week intervention period (ß = –2.30, t(64) = –3.20, p < .01), but no concurrent effects on anxiety (ß = –0.05, t(62) = –0.11, p > .90). Men in CBSM reported significant reductions in depressed mood (ß = –1.21, t(64) = –2.83, p 
.01) from a predicted mean of 14.8 (SE = 1.4) at baseline to a predicted mean of 11.2 (SE = 1.6) immediately following the 10-week intervention. Controls reported marginally significant increases (ß = 1.10, t(64) = 1.90, p = .06) from a predicted mean of 10.7 (SE = 1.8) at baseline to a predicted mean of 14.0 (SE = 2.1) immediately following the 10-week intervention period. Intervention-related reductions in depressed mood were not maintained through 15 months postrandomization (ß = –0.36, t(149) = –1.34, p 
.10). Observed values for POMS measures across the 15-month investigation period are reported in Table 2.
We observed no intervention-related changes in BDI scores (ß = –0.48, t(65) = –0.98, p > .30) during the 10-week training period. We did note, however, that as a group, only men in CBSM + MAT reported mean BDI values at 10 weeks that were below the clinically significant level (i.e., BDI < 10). Next, we examined the modified BDI-affective composite score (composed of sadness, hopeless, dissatisfaction, and guilt) to cross-validate observed intervention effects on depressed mood. We observed a nearly significant group x time interaction (ß = –0.20, t(64) = –1.95, p = .056) over the 10-week intervention period. Men in CBSM + MAT reported significant reductions in these depressive symptoms (ß = –0.16, t(64) = –2.65, p = .01), while those in MAT-Only reported no changes (ß = 0.04, t(64) = 0.51, p > 0.10). Interestingly, it appears that the symptom showing the greatest group x time change during the 10 weeks was hopelessness (ß = –0.11, t(65) = –2.27, p < .05). Men in CBSM + MAT reported significant reductions in hopelessness (ß = –0.07, t(65) = –2.54, p = .013), while those in MAT-Only reported no changes (ß = 0.04, t(65) = 0.96, p > .10). Observed values for these mood measures across the 15-month investigation period are reported in Table 2.
Structural equation modeling was used to determine if changes in depressed mood during the intervention period could explain decreases in viral load over and above the influence of medication adherence. Because we did not observe intervention-related effects on medication adherence, we were unable to examine this as a potential comediator. However, bearing in mind the potent effects of even mild fluctuations in adherence on HIV viral replication, we included this as an important covariate in the model. The final model (see Figure 3) fit adequately (2 [38] = 47.41, p = .14; CFI = 0.94; RMSEA = 0.050), and men in CBSM + MAT reported significant reductions in POMS-Depression over the 10-weeks (ß = –7.44, t = –3.48, p < .01), which were in turn associated with decreases in HIV viral load over the 15-month investigation (ß = 0.002, t = 2.50, p < .05). For each 1-point reduction in POMS-Depression during the 10-week intervention period, we observed a drop of 0.03 log10 viral load units over the 15 months. Because men in CBSM + MAT experienced a 3.6-point decrease in depressed mood on average, reductions in depressed mood may account for an average decrease of 0.11 log10 viral load units. After setting the previously significant path from group membership to the slope of HIV viral load at zero, the decrement in model fit was nonsignificant (
2 [1] = 0.46, p > .10), consistent with the hypothesis that intervention-related reductions in depressed mood explained the effect of CBSM + MAT on HIV viral load over the 15-month period. Interestingly, the latent variable of adherence was unrelated to change in HIV viral load over the investigation (p > .10) but was significantly related to HIV viral load at 15 months postrandomization (ß = –1.13, t = –2.61, p < .05). Thus, the maintenance of stable medication adherence was important, and this was evident across study conditions. The reduction in depressed mood during CBSM + MAT, however, contributed to further decrements in viral load above and beyond sustained adherence over time.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Just how reductions in depressed mood influence HIV viral load remains unclear. The association between depressed mood and viral load seemed to be independent of changes in medication adherence, due in large part to the relatively high levels of antiretroviral medication adherence (approximately 90%) reported by all participants throughout investigation period. It is also noteworthy that we determined that self-reported adherence (i.e., ACTG) as compared with electronic monitored adherence was a superior predictor of HIV viral load in this cohort of men who have sex with men. These data highlight, as other investigators have noted, that each measure of adherence possesses relative strengths and limitations (26,35). Anecdotally, men in the present study reported difficulties using the MEMS caps because medication bottles provided by the study were perceived as bulky, made it clear to others that participants were taking medications, and did not easily fit into their medication routines. It should also be noted that the MEMS caps were only used on a single medication in a multimedication regimen, thus explaining in part the lack of association between the two adherence indicators in the present study. It may be that the validity of specific adherence measures varies as a function of the medication regimen, as well as the population being examined. Future investigations should seek to examine the differential validity of MEMS and ACTG in single-dose antiretroviral regimens (e.g., Trizivir) among diverse cohorts of HIV-positive persons.
The relationship between decreases in depressed mood and slower HIV disease progression may be explained, in part, by alterations in hypothalamic-pituitary-adrenal axis functioning. In fact, reductions in cortisol have been shown to accompany depression (36) and to covary with reductions in depressed mood during a similar intervention in HIV-positive men (37). A burgeoning literature has also demonstrated that cortisol may influence immune status and health outcomes in HIV-positive persons. Elevated serum cortisol has previously been associated with faster progression to AIDS, increased likelihood of developing an AIDS-related condition, and increased morality over a 9-year period (38). Furthermore, lower 15-hour urinary-free cortisol levels have been related to long-term survival with AIDS (39). Other in vitro data indicate that cortisol decreases maturational selection of CD3+ CD4+ T cells in lymphoid organs (40), synergizes with HIV capsid glycoprotein-120 to enhance rates of cell decline (41), facilitates HIV infection of CD3+ CD4+ T cells (42), and is associated with high rates of apoptosis in CD3+ CD4+ T cells and accessory cells in the lymphoid tissue (43). Thus, neuroendocrine-induced impairments in cellular-immune control of HIV infection are another plausible pathway explaining the effects of depressed mood reductions on HIV viral load.
A major limitation of the present study was the lack of complete follow-up data on many participants. We determined that there were no evident, systematic biases related to whether participants provided follow-up data, but the possibility remains that some unexamined variable differentially influenced follow-up. We were able to partially address this concern by utilizing a mixed-model methodology in which all available data were included in intent-to-treat analyses. However, the relatively low proportion of men returning for follow-up across trials highlights the need for innovative methods to enhance follow-up in behavioral interventions with HIV-positive persons (24). An additional limitation is that intervention-related effects were observed only in a subsample of men with detectable viral load at study entry. Although it is likely that HIV replication continues in men with an undetectable viral load (particularly in latent reservoirs such as the testes), the sensitivity of current immunologic measures may not allow for adequate detection of fluctuations in low-range viral load values for these participants. Consequently, it may be unrealistic to expect behavioral interventions to produce meaningful viral activity changes in the subgroup of HIV+ men with an undetectable viral load. Practically, these individuals may not require as intensive of an intervention in order to maintain optimal viral load levels; they appeared to fare well with MAT only.
Because of the difference in contact time between the two study conditions, we cannot ascertain whether group differences in outcomes are attributable to the inclusion of stress management techniques or differential attention and/or social support; this should be explored in future work. Although we were unable to account for group differences in viral load changes as a function of medication adherence, there remains the possibility that these groups differed in other health behaviors such as substance use and unprotected sexual behaviors, which could affect viral load. This is supported, in part, by the results of previous investigations which have determined that increased cocaine use mediated the association between baseline depression and increases in HIV viral load over a 2-year period (44).
Overall, the finding that a 10-week CBSM + MAT intervention can reduce HIV viral load in men with detectable plasma levels by more than 0.5 log10 units is clinically interesting. In order to replicate these preliminary findings, there is a need for a multisite trial to examine the efficacy CBSM + MAT in a large sample of HIV-positive men who have sex with men, who present with a detectable viral load. Future trials should also incorporate a no-treatment control group to determine if MAT alone is efficacious. Studying a cohort of HIV-positive persons with a history of nonadherence may be helpful in order to examine the efficacy of both CBSM + MAT and MAT-Only in the most vulnerable clinical population. Other relevant HIV-positive populations for similar efficacy trials include women, those with clinical depression, and those with a history of substance dependence. The initiation of HAART in more recently diagnosed populations may be a crucial window of opportunity to begin behavioral interventions, at a point before drug resistance develops and when behavior patterns may be amenable to change.
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
2This was similar to the pattern of results observed when analyses of CBSM + MAT effects on HIV viral load were restricted to study completers. We observed a marginally significant group x time interaction (p = .080) and a marginally significant reduction in the CBSM + MAT group (p = .088).
This research was supported by National Institute of Mental Health grants P01 MH49548 and T32 MH18917.
DOI:10.1097/01.psy.0000195749.60049.63
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  T. B. Baker, R. M. McFall, and V. Shoham Current Status and Future Prospects of Clinical Psychology: Toward a Scientifically Principled Approach to Mental and Behavioral Health Care Psychological Science in the Public Interest, November 1, 2008; 9(2): 67 - 103. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Ironson and H. Hayward Do Positive Psychosocial Factors Predict Disease Progression in HIV-1? A Review of the Evidence Psychosom Med, June 1, 2008; 70(5): 546 - 554. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. W. Carrico and M. H. Antoni Effects of Psychological Interventions on Neuroendocrine Hormone Regulation and Immune Status in HIV-Positive Persons: A Review of Randomized Controlled Trials Psychosom Med, June 1, 2008; 70(5): 575 - 584. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Gore-Felton and C. Koopman Behavioral Mediation of the Relationship Between Psychosocial Factors and HIV Disease Progression Psychosom Med, June 1, 2008; 70(5): 569 - 574. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Naar-King, P. Lam, B. Wang, K. Wright, J. T. Parsons, and M. A. Frey Brief Report: Maintenance of Effects of Motivational Enhancement Therapy to Improve Risk Behaviors and HIV-related Health in a Randomized Controlled Trial of Youth Living with HIV J. Pediatr. Psychol., May 1, 2008; 33(4): 441 - 445. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. H. Novack, O. Cameron, E. Epel, R. Ader, S. R. Waldstein, S. Levenstein, M. H. Antoni, and A. R. Wainer Psychosomatic Medicine: The Scientific Foundation of the Biopsychosocial Model Acad Psychiatry, October 1, 2007; 31(5): 388 - 401. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. W. Carrico, M. O. Johnson, J. T. Moskowitz, T. B. Neilands, S. F. Morin, E. D. Charlebois, W. T. Steward, R. H. Remien, F. L. Wong, M. J. Rotheram-Borus, et al. Affect Regulation, Stimulant Use, and Viral Load Among HIV-Positive Persons on Anti-Retroviral Therapy Psychosom Med, October 1, 2007; 69(8): 785 - 792. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
