The number of HIV-infected patients on highly active antiretroviral therapy (HAART) in Africa doubled in 2005 alone, with approximately 1 in 6 people who needed therapy receiving it by December 2005.1
Studies conducted in sub-Saharan Africa have shown that viral suppression and improvement in immune status can be achieved among patients with near-perfect adherence to antiretroviral therapy (ART).2-4 It has been observed, however, that prolonged exposure to HAART may result in untoward adverse drug reactions, poor adherence, and the emergence of drug-resistant mutants.5,6
Despite increasing access to antiretroviral drugs, the long-term success of treatment programs in resource-limited settings requires establishing the levels of and reasons for patients' discontinuation or modification of ART. These data, which are scarce at the moment, can potentially provide a long-term strategic approach to initial and subsequent decisions regarding ART.7 In this study, we sought to estimate the prevalence and to identify the factors associated with discontinuation and with modification of HAART among HIV-positive Ugandans.
The study was conducted in 2 treatment centers that provide clinical care to patients with HIV in Uganda; Mulago Hospital and the Joint Clinical Research Centre (JCRC). Mulago Hospital is the national referral hospital, and it provides free ART to approximately 4300 people, most of whom come from the adjacent Infectious Diseases Institute. The JCRC specializes in HIV/AIDS care and research, with a large and diverse cumulative patient population of more than 19,000 clients on ART. Before 2004, most patients at the JCRC clinic paid for their ART, but access to free ART by poor patients in that clinic increasingly became available in 2004. Currently, free therapy at the JCRC is nearly exclusively limited to women (mostly poor women and/or widows), orphans and vulnerable children (less than 18 years of age), and their caretakers.
The study population consisted of consecutively sampled HIV-infected patients 18 years of age and older who were receiving HAART and those who had previously initiated ART but were presently off treatment. Subjects were included in the analysis if they had been stable on HAART for at least 1 month.
HAART was defined as (1) 2 nucleoside reverse transcriptase inhibitors (NRTIs) in combination with at least 1 protease inhibitor (PI) or nonnucleoside reverse transcriptase inhibitor (NNRTI), (2) 1 NRTI in combination with at least 1 PI and 1 NNRTI, or (3) an abacavir- or tenofovir-containing regimen of 3 or more NRTIs in the absence of PIs and NNRTIs. In this study, discontinuation of therapy involved complete cessation of all antiretroviral drugs used in a HAART regimen for a period of at least 1 month. This definition encompassed stopping ART as the decision of the physician or as a result of default by the patients themselves. It excluded structured treatment interruption, where participants stop and restart treatment at predefined intervals, and CD4 cell count-guided structured intermittent treatment. Modification of therapy was defined as the switching or changing of at least 1 antiretroviral used as part of an initial HAART regimen. More than 1 change for an individual was registered as a single modification in this study. Dosage adjustments were not considered to be modifications.
From December 2005 through March 2006, we conducted a cross-sectional study in which patients who gave informed consent received semistructured quantitative and unstructured qualitative face-to-face interviews. We used the public domain of medical and pharmacy records of the previous 2-year period to attempt to trace individuals who had stopped coming to the clinics for their antiretroviral drugs. Ethical approval to conduct this study was obtained from the Makerere University Faculty of Medicine Research and Ethics Committee and the Uganda National Council of Science and Technology.
Data were double entered into Epi Info 6 (Centers for Disease Control and Prevention, Atlanta, GA) and exported to SPSS software (version 12.0; SPSS, Chicago, IL) for statistical analysis. Descriptive statistics were used to summarize characteristics of study patients. The estimated prevalences of HAART discontinuation and HAART modification were reported as percentages. We assessed the association between HAART discontinuation and HAART modification (dichotomized) with each categoric independent variable using χ2 tests or Fisher exact tests. Assessment of normally distributed continuous predictor variables was done using independent t tests, and assessment of skewed continuous predictor variables was done with Wilcoxon rank sum tests. The level of significance was set at P < 0.05. Odds ratios (ORs) and their 95% confidence intervals (CIs) were also computed. All tests of significance were 2-sided. To determine factors that were independently associated with HAART discontinuation and with HAART modification, independent variables with P ≤ 0.2 at the bivariate level were entered into multiple logistic regression models. Factors that were used to model HAART discontinuation included age, monthly income, adherence (self-report), duration on HAART, HAART experience, type of HAART regimen, use of alternative medicines, history of hospitalization, and distance from clinic. Variables used to model HAART modification included age, monthly income, marital status, type of HAART regimen, duration of time on HAART, use of alternative medicines, history of hospitalization, opportunistic infection, and most recent CD4 cell count. The backward elimination method (likelihood ratio test) was used to determine the final multivariate models, with a removal level of significance of P < 0.10. Sensitivity of results was assessed for consistency using the forward selection method.
A total of 686 participants were interviewed. Of these, 15 (2.2%) were excluded from the analysis because they had been on HAART for less than 1 month and were unstable on therapy. As seen in Table 1, two thirds of participants were female (65.9%). The median age was 36 years (interquartile range [IQR]: 31-43 years), whereas the median CD4 count (n = 478) at last measurement was 175 cells/μL (IQR: 66-298 cells/μL). The largest proportion of respondents (83.8%) received NNRTI-based regimens.
Ninety-four (13.7%) patients had ever discontinued therapy, whereas 175 (25.5%) had ever modified their initial HAART regimen (Table 2). The most frequently cited reason for discontinuing therapy was that drugs were too expensive (43%). The second most commonly cited reason was to avoid side effects and toxicity (21.1%). The most common reasons for modifying therapy were to avoid side effects (71.8%) and high drug cost (23.3%).
Factors Associated With Discontinuation of Therapy
Table 3 shows the factors that were independently associated with discontinuation of HAART in the final binary logistic regression model. Individuals with ART experience before starting their current regimen were almost 4 times (OR = 3.70, 95% CI: 2.13 to 6.25; P < 0.001) as likely to discontinue therapy compared with those who were naive to ART at the time of starting their present regimen. Other independently associated factors included hospitalization after starting ART (OR = 2.36, 95% CI: 1.32 to 4.20; P = 0.004), use of alternative medicines (OR = 2.18, 95% CI: 1.06 to 4.47; P = 0.033), and receiving HAART for 1 year or less (vs. more than 1 year; OR = 11.11, 95% CI: 5.00 to 25.00; P < 0.001). Subgroup analyses of patients on NNRTI-based regimens and defaulters as the only discontinuers yielded similar results (data not shown). We had sufficient power for both sensitivity analyses.
Factors Associated With Modification of Therapy
On multivariate analysis, factors independently associated with modification of HAART (Table 4) included being unmarried (OR = 1.64, 95% CI: 1.02 to 2.70; P = 0.042) and more than 3 months' duration on therapy (OR = 3.13, 95% CI: 1.16 to 8.33; P = 0.041).
This cross-sectional study has shown that 13.7% of the study participants had ever stopped taking all their ART drugs for at least 1 month, whereas 25.5% changed 1 or more components of their initial HAART regimen(s). These relatively high proportions are consistent with results obtained in previous studies.7-12 Exclusion of patients who had not yet completed 1 month on therapy did not alter these estimates. Such patients started first-line (nevirapine-based) therapy that required dose escalation for the first 14 days of treatment, and were therefore not yet stable on ART. It is important to note that almost half (45.7%) of the discontinuations were initiated by the patients (see Table 2), and similar observations have been reported elsewhere.10,13,14 Although it may be inevitable to avoid unplanned treatment interruptions in clinical practice, complete cessation of ART poses a major challenge in our resource-limited setting. The first-line regimen in Uganda has mainly been rolled out as a generic fixed-dose combination of stavudine, lamivudine, and nevirapine.15 Stopping this regimen requires that separate components of stavudine and lamivudine be administered for an additional 1 week to compensate for the long half-life of nevirapine.16 This is not practicable with the fixed-dose combinations currently available. As such, nevirapine resistance resulting from subtherapeutic plasma drug levels is bound to develop,17 which would limit future treatment options of NNRTI-based regimens.18
In this study, most participants (88.5%) were receiving free antiretroviral drugs (data not shown). The fact that drug cost was the most frequently cited reason for discontinuation of ART suggests that respondents may have interrupted treatment because of financial constraints in the era before provision of free therapy. Drug cost has been known to be a major barrier to utilization of ART in resource-limited settings.19 It is interesting to note that the percentage of discontinuation among individuals on free therapy seemed to be higher than that of individuals on fee-for-service ART (13.8% [84 of 607 participants] vs. 12.7% [10 of 79 participants]), although this difference was not statistically significant (P = 0.77), given the small numbers of participants who paid for ART (data not shown). Qualitative findings revealed that individuals who had firm financial backing (eg, those whose employers or financially stable relatives paid for their antiretroviral drugs) continued to purchase their drugs and were less likely to interrupt therapy. The poorer individuals probably accounted for the higher proportion of discontinuation among persons on free ART, because it is logical that they crossed over from paid drugs after interrupting therapy for reasons related to drug cost. At the time of this study, however, monthly income was not independently associated with discontinuation of therapy, suggesting that international donor programs and government ART roll-out programs may have played an important role in removing cost as a barrier to treatment access.9 That notwithstanding, ART programs that charge fees ought to be replaced by those that offer free antiretroviral drugs. This should reduce the burden of ART discontinuation resulting from inability of individuals to afford antiretroviral drugs.
Consistent with other findings,13,20-22 we observed that a shorter duration on ART was independently associated with discontinuation of therapy. In the present study, this result was interpreted as suggesting that individuals who had just started ART were not well adjusted on treatment and were therefore more likely to experience interruptions attributable to early drug toxicities. Another probable reason was that patients who had taken treatment much longer received ongoing information that helped them to adhere better.
Patients who used alternative medicines were more than twice as likely to discontinue HAART compared with those who did not. It should be noted, however, that most (79%) of those who reported using alternative medicines specifically used traditional herbal treatments (see Table 1). Qualitative findings revealed that such patients obtained herbal treatments from traditional healers, which they then took to augment ART. This contrasts with results from the developed world, where conventional multivitamin and mineral dietary supplements are the most commonly used complementary and alternative medicines.23-26 We also found out that some patients stopped ART altogether and resorted to the use of local herbs. Unfortunately, clients concealed this information from health care providers at the treatment clinics. It was also reported in the qualitative interviews that some traditional healers combined their local herbs with antiretroviral drugs. These findings present a challenge to the utilization of ART in our resource-limited setting. It is estimated that more than 70% of the African population seek their first line of care from traditional health practitioners.27 The unique position that traditional health practitioners occupy can therefore be exploited to refer HIV-infected individuals for routine treatment and evaluation. Traditional healers should then be sensitized about ART to prevent likely interactions between herbal treatments and antiretroviral drugs,28 uphold acceptable standards of care, and evaluate the use of herbal treatments in the management of AIDS-defining illnesses.29
Respondents who had prior exposure to ART before starting their current regimen were almost 4 times as likely to have ever stopped ART compared with those who were previously naive to antiretroviral drugs before starting their current regimen. Qualitative findings revealed that patients starting ART for the first time in this setting are usually quite sick and realize the need to take their drugs. To corroborate this finding, the Infectious Diseases Institute reported a mean baseline CD4 count of 63 cells/μL at ART initiation.30 Conversely, treatment-experienced patients may have underlying drug resistance arising from prior inadequate exposure to antiretroviral agents,31 resulting in earlier interruption of treatment.
Modification of HAART was associated with marital status (unmarried) and longer duration on ART. Qualitatively, it was observed that couples who disclosed to each other and received counseling from the same health facility were likely to adhere better. This observation is in agreement with the findings of Waddell and Messeri,32 who reported that social support after disclosure enhanced use of combination ART. This form of social support, which was protective of HAART modification among married couples, was probably lacking in the unmarried individuals.
Duration on therapy for more than 3 months was associated with a higher likelihood of modifying ART, whereas a prospective study in India9 estimated a median switching time of 13 months. This might be interpreted as suggesting that the longer individuals are exposed to antiretroviral drugs, the more likely they are to experience poor adherence,6 long-term adverse drug effects,20 and treatment failure.9
This study has several strengths. First, it distinguished between discontinuation and modification of HAART, which most previous studies had not. Second, we recruited patients who were ART naive and those who were ART experienced at the time of initiating their current HAART regimen, thus reflecting what happens in “real-life” clinical practice. Previous studies have focused on treatment-naive patients, who are known to have higher adherence to therapy. Third, the patients served by the 2 treatment centers at the time of data collection came from all over the country. The JCRC was the first HIV/AIDS clinic in the country, whereas Mulago Hospital is a national referral center, where scale-up of ART was largely pioneered. Our findings are therefore generalizable. This observational study has important limitations, however. First, recall bias may have influenced these results, because self-report was used as the main method of inquiry. Use of document reviews to validate self-report was not possible. It was observed that several patients had transferred from their initial centers of ART provision to the present centers. As such, the latter centers did not have documentation of the earlier medical and/or pharmacy records of these patients. For those who initiated ART at the present centers, data were incomplete. Second, the cross-sectional nature of this study may not allow for a direct investigation of causal relations between the factors studied and the outcomes of interest, because we could not establish, at measurement, the temporal sequence of the dependent and independent factors. For some factors (HAART experience, use of alternative medicines, and duration on HAART), however, their historical time sequence with the outcomes was known or established. Third, discontinuation of ART may have been underestimated, because some individuals who were not presently on therapy and who had stopped coming to the treatment centers could not be traced. From the records of the previous 2-year period, estimates of persons who completely stopped coming to the 2 sites for ART because they were deceased or lost to follow-up ranged from 5% to 6%. This may have been offset by the possibility that individuals who discontinued in other centers where treatment was on a fee-for-service basis may have enrolled for free therapy at the health facilities we studied, however. Fourth, we measured total time of individuals on ART and not the actual time from treatment initiation to discontinuation of therapy. As such, a potential bias toward the null may have arisen, because we only estimated the overall risk of treatment interruption in relation to total time since initiation of ART rather than the risk associated with the first discontinuation of therapy. The approach we used minimized patient recall, however.
In conclusion, the proportions of subjects who discontinue or modify HAART in our resource-constrained setting present a challenge to the limited treatment options that we currently have. Furthermore, cost of antiretroviral drugs as the most frequent reason for discontinuation of HAART has major implications for ART programs that charge fees in resource-limited settings. The factors associated with discontinuation and modification of HAART observed in this cross-sectional study should be investigated further in longitudinal studies of ART utilization.
The authors acknowledge the help of Herbert Kayiga, Richard Kagalama, Jeff Ssenyonjo, Emmanuel Kusasira, and Fred Nsubuga in collecting the data. They also thank the study participants from whom these data were obtained. Special thanks go to the staff of Mulago Hospital Central Pharmacy and those of the JCRC, whose invaluable input into this project has made it a success.
2. Coetzee D, Boulle A, Hildebrand K, et al. Promoting adherence to antiretroviral therapy: the experience from a primary care setting in Khayelitsha, South Africa. AIDS
. 2004;18(Suppl 3):S27-S31.
3. Weidle PJ, Malamba S, Mwebaze R, et al. Assessment of a pilot antiretroviral drug therapy programme in Uganda: patients' response, survival, and drug resistance. Lancet
4. Wools-Kaloustian K, Kimaiyo S. Extending HIV care in resource-limited settings. Curr HIV/AIDS Rep
5. Carr A, Cooper DA. Adverse effects of antiretroviral therapy. Lancet
6. Paterson DL, Swindells S, Mohr J, et al. Adherence to protease inhibitor therapy and outcomes in patients with HIV infection. Ann Intern Med
7. Kirstein LM, Greenblatt RM, Anastos K, et al. Prevalence and correlates of highly active antiretroviral therapy switching in the Women's Interagency HIV Study. J Acquir Immune Defic Syndr
8. Grant LA, Silverberg MJ, Palacio H, et al. Discontinuation of potent antiretroviral therapy: predictive value of and impact on CD4 cell counts and HIV RNA levels. AIDS
9. Kumarasamy N, Vallabhaneni S, Cecelia AJ, et al. Reasons for modification of generic highly active antiretroviral therapeutic regimens among patients in southern India. J Acquir Immune Defic Syndr
10. Mocroft A, Phillips AN, Soriano V, et al. Reasons for stopping antiretrovirals used in an initial highly active antiretroviral regimen: increased incidence of stopping due to toxicity or patient/physician choice in patients with hepatitis C coinfection. AIDS Res Hum Retroviruses
11. Mocroft A, Youle M, Moore A, et al. Reasons for modification and discontinuation of antiretrovirals: results from a single treatment centre. AIDS
12. Ripamonti D, Arici C, Pezzotti P, et al. Hepatitis C infection increases the risk of the modification of first highly active antiretroviral therapy in HIV-infected patients. AIDS
13. d'Arminio Monforte A, Cozzi-Lepri A, Phillips A, et al. Interruption of highly active antiretroviral therapy in HIV clinical practice: results from the Italian Cohort of Antiretroviral-Naive Patients. J Acquir Immune Defic Syndr
14. O'Brien ME, Clark RA, Besch CL, et al. Patterns and correlates of discontinuation of the initial HAART regimen in an urban outpatient cohort. J Acquir Immune Defic Syndr
15. Colebunders R, Kamya MR, Laurence J, et al. First-line antiretroviral therapy in Africa-how evidence-based are our recommendations? AIDS Rev
16. Mackie NE, Fidler S, Tamm N, et al. Clinical implications of stopping nevirapine-based antiretroviral therapy: relative pharmacokinetics and avoidance of drug resistance. HIV Med
17. Flys TS, Chen S, Jones DC, et al. Quantitative analysis of HIV-1 variants with the K103N resistance mutation after single-dose nevirapine in women with HIV-1 subtypes A, C, and D. J Acquir Immune Defic Syndr
18. Jourdain G, Ngo-Giang-Huong N, Le Coeur S, et al. Intrapartum exposure to nevirapine and subsequent maternal responses to nevirapine-based antiretroviral therapy. N Engl J Med
19. Byakika-Tusiime J, Oyugi JH, Tumwikirize WA, et al. Adherence to HIV antiretroviral therapy in HIV+ Ugandan patients purchasing therapy. Int J STD AIDS
20. Bongiovanni M, Bini T, Capetti A, et al. Long-term antiretroviral efficacy and safety of lopinavir/ritonavir in HAART-experienced subjects: 4 year follow-up study. AIDS
21. d'Arminio Monforte A, Lepri AC, Rezza G, et al. Insights into the reasons for discontinuation of the first highly active antiretroviral therapy (HAART) regimen in a cohort of antiretroviral naive patients. I.CO.N.A. Study Group. Italian Cohort of Antiretroviral-Naive Patients. AIDS
22. Li X, Margolick JB, Conover CS, et al. Interruption and discontinuation of highly active antiretroviral therapy in the Multicenter AIDS Cohort Study. J Acquir Immune Defic Syndr
23. Cho M, Ye X, Dobs A, et al. Prevalence of complementary and alternative medicine use among HIV patients for perceived lipodystrophy. J Altern Complement Med
24. Dhalla S, Chan KJ, Montaner JS, et al. Complementary and alternative medicine use in British Columbia-a survey of HIV positive people on antiretroviral therapy. Complement Ther Clin Pract
25. Mikhail IS, DiClemente R, Person S, et al. Association of complementary and alternative medicines with HIV clinical disease among a cohort of women living with HIV/AIDS. J Acquir Immune Defic Syndr
26. Standish LJ, Greene KB, Bain S, et al. Alternative medicine use in HIV-positive men and women: demographics, utilization patterns and health status. AIDS Care
28. Homsy J, King R, Balaba D, et al. Traditional health practitioners are key to scaling up comprehensive care for HIV/AIDS in sub-Saharan Africa. AIDS
29. Homsy J, Katabira E, Kabatesi D, et al. Evaluating herbal medicine for the management of herpes zoster in HIV-infected patients in Kampala, Uganda. Journal of Complementary Medicine
30. Colebunders R, Ronald A, Katabira E, et al. Rolling out antiretrovirals in Africa: there are still challenges ahead. Clin Infect Dis
31. Rodes B, Garcia F, Gutierrez C, et al. Impact of drug resistance genotypes on CD4+ counts and plasma viremia in heavily antiretroviral-experienced HIV-infected patients. J Med Virol
32. Waddell EN, Messeri PA. Social support, disclosure, and use of antiretroviral therapy. AIDS Behav