Multiple studies support initiation of HIV antiretroviral (ARV) therapy (ART) in healthy individuals at CD4 cell counts above the threshold of risk for opportunistic infections.1–4 Early initiation of ARV treatment may have substantial public health benefits by reducing the risk of HIV transmission in discordant couples.5 These findings have provided increased support for universal testing with immediate ART initiation to prevent both HIV-related morbidity and HIV transmission.6 Recommending treatment for HIV-infected individuals regardless of CD4 cell count will necessitate offering therapy to increasing numbers of asymptomatic patients, particularly in sub-Saharan Africa, where a large proportion of the HIV-infected population is untreated.7
Although adherence to ARVs in sub-Saharan Africa has been excellent in general,8 most adherence studies have been limited to people with advanced disease. Advanced disease has significant functional and economic impacts on the individual and their family.9 ARV treatment adherence in resource-limited settings is sustained, in part, by tangible support to overcome economic barriers to sustained treatment access10–12 and is reinforced by functional recovery and subsequent reversal of household economic strains incurred by caring for someone with advanced disease.13,14 Thus, individuals initiating ARVs at higher CD4 cell counts in these settings might lack elements of social support that sustain early adherence. We examined whether treatment initiation at higher CD4 cell counts is associated with lower adherence and viral suppression in a population of patients initiating ART in rural Uganda.
Study Methods and Patient Population
We performed a prospective observational study of HIV-infected individuals enrolled from a public hospital in southwestern Uganda. Study participants were recruited from the Mbarara Regional Referral Hospital Immune Suppression Syndrome Clinic, which dispenses free ART in the region. Patients older than 18 years who were initiating ARVs and lived within 60 km from the clinic were eligible for study participation. The study was approved by the Institutional Review Committees of Mbarara University of Science and Technology, Partners Healthcare, and the University of California, San Francisco. All participants gave written informed consent.
At the enrollment visit, we collected demographic data including age, marital status, educational attainment, socioeconomic status, self-reported distance from clinic (in minutes of travel time), self-reported physical functioning [Medical Outcomes Study Physical Health Summary (MOS-PHS) Score15], screen for heavy drinking [3-item consumption subset of the Alcohol Use Disorders Identification Test (AUDIT-C)],16 and depression symptom severity (15-item Hopkins Symptom Checklist for Depression, modified for the local context with the addition of a 16th item, “feeling like I don’t care about my health”).17 Blood was collected for HIV-RNA and CD4 cell count at baseline and again at 3 months. We included participants who had a repeat viral load test within 120 days in the analysis of virologic outcomes. Participants who did not return for a second visit by 120 days were considered loss to follow-up. CD4 cell count, our primary exposure of interest, was dichotomized at the threshold of 250 cells/μL.
ARV adherence was measured using MEMSCap pill bottles (Aardex, Switzerland), which electronically record the date and time of pill bottle opening. Participants were visited at home once monthly, and MEMs data were downloaded. Because physical function changes quickly with the initiation of ART,18 we focused on adherence in the first 90 days as the most sensitive interval to estimate the impact of initial stage of disease on adherence. Our primary outcomes were (a) average adherence in the first 90 days of therapy of less than 90%, (b) any treatment interruptions (defined as zero adherence for >72 hours continuously) in the first 90 days of therapy; (c) number of treatment interruptions in the first 90 days of therapy; and (d) persistent detectable HIV viremia at 90 days. We selected a duration of 72 hours based on previous data supporting it as a threshold required to detect viral replication.19
We compared baseline characteristics between the two exposure groups (CD4 cell count <250 and ≥250 cells/μL) using χ2 testing for categorical variables and nonparametric rank-sum testing for continuous, non-normally distributed variables. For binomial outcomes (adherence < 90%, any treatment interruption, and persistent viremia), we fit logistic regression models to estimate their associations with our primary explanatory variable of interest, baseline CD4 cell count ≥250 vs <250 cell/μL. For number of treatment interruptions, we fit a negative binomial regression model to estimate the incidence rate ratio comparing those with CD4 cell counts ≥250 vs <250 cell/μL. We used univariable and multivariable regression modeling to identify potential predictors of adherence including age, sex, marital status, educational attainment (greater than primary education vs primary education or none), employment, socioeconomic status as measured by the Filmer–Pritchett Asset Index,20 self-reported distance to clinic (>60 minutes of travel time vs 60 minutes or less), depression symptom severity, positive screen for heavy drinking, and ARV dosing frequency (daily vs more than daily). Because adherence monitoring was censored at the time of death or loss to follow-up, we also repeated analyses allocating those with missing data to poor outcome groups (average adherence < 90%, occurrence of at least one treatment gap, and detectable viremia) to assess for potential bias from missing data.
We included 473 participants in the analysis. The majority of participants was women (70.6%) and had a median age of 34 years [interquartile range (IQR), 29–39 years]. The median CD4 cell count was 132 cells/μL (IQR, 16–200 cells/μL), and the median HIV viral load (log10) was 5.0 copies/mL (IQR, 1.6–7.0 copies/mL), (Table 1). The MOS-PHS summary score was higher in those with CD4 cell counts ≥ 250 cells/μL (median 45 vs 40, P = 0.01). In the 120-day period after initiation, there were 2 deaths (3.3%) among participants with a CD4 cell count ≥ 250 cells/μL and 15 deaths (3.6%) among participants with a CD4 cell count < 250 cells/μL. Another 5 (8.6%) and 26 (6.5%) participants in each group, respectively, were lost to follow-up. Average adherence in the first 90 days was 89.0% across all participants and did not differ by baseline CD4 cell count (P = 0.08). Fifty-two participants (10.9%) had at least 1 treatment interruption of 72 hours, and the median duration of interruption was 8 days (IQR, 8–18 days). There were approximately twice as many individuals with a CD4 cell count ≥ 250 cells/μL who had any treatment interruptions compared with those with <250 CD4 cells/μL (20.0% vs 9.7%) (P = 0.02). Three hundred fifty-eight participants (75.7%) had a repeat viral load test within 120 days and were included in the analysis of persistent viremia. The proportion of patients with viral suppression on follow-up testing was 19% vs 9.5%; (P = 0.06). There was no difference between groups in the level of viremia upon repeat testing [median 2.2 (log10) copies/mL in both groups, P = 0.45].
In multivariable regression models, a CD4 cell count ≥ 250 was associated with increased odds of any treatment interruption [adjusted odds ratio (AOR), 2.28; 95% confidence interval (CI), 1.01 to 5.15; P = 0.048], increased number of treatment interruptions [adjusted incidence rate ratio (AIRR), 2.56; 95% CI, 0.99 to 6.65; P = 0.054], and increased odds of persistent detectable viremia (AOR, 2.83; 95% CI, 1.14 to 7.00; P = 0.024) (Table 2). CD4 cell count at baseline was not associated with average adherence < 90% (AOR, 1.40; 95% CI, 0.70 to 2.82; P = 0.344). In analyses allocating participants who had died or were lost to follow-up to poor outcomes, we estimated an attenuated association for odds of treatment interruptions (AOR, 1.78; 95% CI, 0.88 to 3.60; P = 0.108) but a persistent effect for detectable viremia (AOR, 2.65; 95% CI, 1.08 to 6.52; P = 0.034). The association with average adherence < 90% was unchanged (AOR, 1.60; 95% CI, 0.81 to 3.18; P = 0.179).
We found that individuals starting ARVs at a CD4 cell count ≥ 250 cells/μL in southwestern Uganda were more likely to experience treatment interruptions in the first 3 months of therapy and were more likely to have persistent viremia at 3 months, compared with those starting with a CD4 cell count < 250 cells/μL. This finding was despite lower baseline viral load in those initiating therapy with higher CD4 cell counts. Although ARV adherence has generally been excellent in resource-limited settings,8,21 these estimates have been based on studies of people with advanced disease, which can have a profound impact on both individuals and their social networks. Social support helps HIV-infected persons overcome severe structural and economic barriers to sustain adherence, which in turn restores their health and economic contributions to the social network.11,22 Relatively healthy, asymptomatic individuals taking ARVs may face the same level of structural and economic barriers, but their treatment course is not characterized by the same changes in functional status as is frequently observed among individuals with advanced disease initiating treatment. Consequently, those initiating treatment may be less likely to engage the commitment of their social network to overcome structural barriers to treatment adherence.22
Our findings are consistent with variable ARV adherence observed among healthy HIV-infected and HIV-uninfected individuals taking ARVs in the form of preexposure prophylaxis (PrEP). As few as 50% of these individuals have detectable drug levels when ARVs are prescribed for preventive purposes,23–25 although adherence to PrEP is higher when HIV-infected sexual partners are included in interventions to increase the provision of support to the HIV-uninfected partner.26,27 ARV treatment for asymptomatic HIV-infected persons might be analogous in some respects to PrEP, in that it is given to prevent adverse outcomes rather than to restore health. An important distinction between our study and the others described is our use of an observational design (vs an intervention trial), which might be more representative of adherence in similar settings.
Although baseline CD4 cell count was associated with treatment interruptions, it was not associated with average adherence. Treatment interruptions particularly predispose individuals to development of resistance to non-nucleoside reverse transcriptase inhibitors,19,21,28 which continue to be first-line therapy in most resource-limited settings.
Our study has several limitations. First, we were not able to characterize the reasons why participants initiated ART. Those with a CD4 cell count ≥ 250 cells/μL had higher MOS-PHS scores but might not have had improved health status. Some might have been initiated on ARVs for other health events or comorbidities not captured in the MOS-PHS score. Although a limitation, unmeasured illness or clinical deterioration in the higher CD4 cell count group would bias the findings to the null. Higher CD4 cell counts, however, might have been associated with other, unmeasured confounders that explain the association we found. A related limitation is that the exposure of interest (high vs low CD4 cell count) was not assigned randomly. In the absence of an experimental study design, we are unable to assert that the observed associations are causal. Third, we had relatively few people in the ≥250-cells/μL group (n = 60), which led to wide CIs in our estimated effect sizes.
Additional investigation of ARV adherence in relatively healthy individuals in resource-limited settings is recommended to confirm our findings. Moreover, elucidating the factors that cause poor adherence will be vital to developing targeted interventions to improve outcomes. Although a variety of interventions, including pre- and intratherapy adherence counseling29,30 and automated medication reminders,31 have proven effective, these studies did not target those with relatively high CD4 counts. The efficacy of these and other interventions specifically targeted to relatively healthy individuals should be pursued as ART is made available to those at higher CD4 count thresholds. Although ARVs have great potential to improve health among those infected and to prevent transmission to those uninfected,1–3,5,23 healthy individuals receiving ARVs might warrant additional adherence support.
The authors thank the UARTO participants who made this study possible and Annet Kembabazi, Annet Kawuma, and Anna Baylor, for providing study coordination and support.
1. Kitahata MM, Gange SJ, Abraham AG, et al.. Effect of early versus deferred antiretroviral therapy for HIV on survival. N Engl J Med. 2009;360:1815–1826.
2. Severe P, Juste MA, Ambroise A, et al.. Early versus standard antiretroviral therapy for HIV-infected adults in Haiti. N Engl J Med. 2010;363:257–265.
3. Siegfried N, Uthman OA, Rutherford GW. Optimal time for initiation of antiretroviral therapy in asymptomatic, HIV-infected, treatment-naive adults. Cochrane Database Syst Rev. 2010:CD008272.
4. Sterne JA, May M, Costagliola D, et al.. Timing of initiation of antiretroviral therapy in AIDS-free HIV-1-infected patients: a collaborative analysis of 18 HIV cohort studies. Lancet. 2009;373:1352–1363.
5. Cohen MS, Chen YQ, McCauley M, et al.. Prevention of HIV-1 infection with early antiretroviral therapy. N Engl J Med. 2011;365:493–505.
6. Granich RM, Gilks CF, Dye C, et al.. Universal voluntary HIV testing with immediate antiretroviral therapy as a strategy for elimination of HIV transmission: a mathematical model. Lancet. 2009;373:48–57.
8. Mills EJ, Nachega JB, Buchan I, et al.. Adherence to antiretroviral therapy in sub-Saharan Africa and North America: a meta-analysis. JAMA. 2006;296:679–690.
9. Yamano T, Jayne TS. Measuring the impacts of working-age adult mortality on small-scale farm households in Kenya. World Dev. 2004;32:91–119.
10. Nabyonga-Orem J, Bazeyo W, Okema A, et al.. Effect of HIV/AIDS on household welfare in Uganda rural communities: a review. East Afr Med J. 2008;85:187–196.
11. Ware NC, Idoko J, Kaaya S, et al.. Explaining adherence success in sub-Saharan Africa: an ethnographic study. PLoS Med. 2009;6:e11.
12. Baylies C. The impact of AIDS on rural households in Africa: a shock like any other? Dev Change. 2002;33:611–632.
13. Thirumurthy H, Zivin JG, Goldstein M. The economic impact of AIDS treatment labor supply in Western Kenya. J Hum Resour. 2008;43:511–552.
14. Thirumurthy H, Jafri A, Srinivas G, et al.. Two-year impacts on employment and income among adults receiving antiretroviral therapy in Tamil Nadu, India: a cohort study. AIDS. 2011;25:239–246.
15. Revicki DA, Sorensen S, Wu AW. Reliability and validity of physical and mental health summary scores from the Medical Outcomes Study HIV Health Survey. Med Care. 1998;36:126.
16. Bush K, Kivlahan DR, McDonell MB, et al.. The AUDIT alcohol consumption questions (AUDIT-C): an effective brief screening test for problem drinking. Ambulatory care quality Improvement Project (ACQUIP). Alcohol Use Disorders Identification test. Arch Intern Med. 1998;158:1789–1795.
17. Bolton P, Ndogoni L. Cross-cultural Assessment of Trauma-related Mental Illness (Phase II): A Report of Research Conducted by World Vision Uganda and the Johns Hopkins University. Baltimore, MD: The Johns Hopkins University; 2001.
18. Stangl A, Wamai N, Mermin J, et al.. Trends and predictors of quality of life among HIV-infected adults taking highly active antiretroviral therapy in rural Uganda. AIDS Care. 2007;19:626–636.
19. Parienti JJ, Ragland K, Lucht F, et al.. Average adherence to boosted protease inhibitor therapy, rather than the pattern of missed doses, as a predictor of HIV RNA replication. Clin Infect Dis. 2010;50:1192–1197.
20. Filmer D, Pritchett LH. Estimating wealth effects without expenditure data—or tears: an application to educational enrollments in states of India. Demography. 2001;38:115–132.
21. Oyugi JH, Byakika-Tusiime J, Charlebois ED, et al.. Multiple validated measures of adherence indicate high levels of adherence to generic HIV antiretroviral therapy in a resource-limited setting. J Acquir Immune Defic Syndr. 2004;36:1100–1102.
22. Tsai AC, Bangsberg DR. The importance of social ties in sustaining medication adherence in resource-limited settings. J Gen Intern Med. 2011;26:1391–1393.
23. Grant RM, Lama JR, Anderson PL, et al.. Preexposure chemoprophylaxis for HIV prevention in men who have sex with men. N Engl J Med. 2010;363:2587–2599.
24. Martin M, Vanichesni S, Suntharasamai P, et al.. Participant adherence in the Bangkok Tenofovir Study, an HIV pre-exposure prophylaxis trial in Bangkok. Paper presented at: 6th International AIDS Society Conference on HIV Pathogenesis, Treatment, and Prevention; July 17–20, 2011; Rome, Italy. [Abstract TUPE350].
25. Van Damme L, Corneli A, Ahmed K, et al.. Preexposure prophylaxis for HIV infection among African women. N Engl J Med. 2012;367:411–422.
26. Haberer J, Baeten J, Celum C, et al.. Near perfect early adherence to antiretroviral PrEP against HIV infection among HIV serodiscordant couples as determined by multiple measures: preliminary data from the partners PrEP study. Paper presented at: 18th Conference on Retroviruses and Opportunistic Infections; February 27 - March 2, 2011; Boston, Massachusetts. Paper #488.
27. Baeten JM, Donnell D, Ndase P, et al.. Antiretroviral prophylaxis for HIV prevention in heterosexual men and women. N Engl J Med. 2012;367:399–410.
28. Genberg BL, Wilson IB, Bangsberg D, et al.. Patterns of ART adherence and impact on HIV RNA among patients from the MACH14 study. AIDS. 2012;26:1415–1423.
29. Chung MH, Richardson BA, Tapia K, et al.. A randomized controlled trial comparing the effects of counseling and alarm device on HAART adherence and virologic outcomes. PLoS Med. 2011;8:e1000422.
30. Pradier C, Bentz L, Spire B, et al.. Efficacy of an educational and counseling intervention on adherence to highly active antiretroviral therapy: French prospective controlled study. HIV Clin Trials. 2003;4:121–131.
31. Lester RT, Ritvo P, Mills EJ, et al.. Effects of a mobile phone short message service on antiretroviral treatment adherence in Kenya (WelTel Kenya1): a randomised trial. Lancet. 2010;376:1838–1845.