Highly active antiretroviral therapy (ART) has transformed HIV infection from a near-uniformly fatal disease into a chronic, manageable illness. However, in contrast to developed countries, ART is unavailable to most of HIV-infected persons in developing countries, where more than 90% of these patients reside. 1
In Brazil a public health program of large-scale, free-of-charge distribution of antiretroviral drugs to all HIV-infected persons who fulfill criteria established by an independent advisory committee was implemented in 1996. At present, this program provides treatment to more than 130,000 Brazilians with HIV infection. Drugs are distributed from more than 400 dispensary units around the country, and treatment is monitored through a national network of laboratories that perform CD4+ lymphocyte counts and HIV RNA plasma viral load measurements. 2,3
In Brazil, as in other parts of the world where ART is used, therapeutic failure is one of the main concerns. 2,3 Factors that influence therapeutic success and failure have been identified in developed countries. 4–6 However, this issue has not been studied in public clinics in a developing country setting. In this study we assessed the response to therapy in a sample of patients receiving antiretroviral drugs in public clinics in Rio de Janeiro, Brazil, and factors associated with failure of therapy.
Study Setting and Data Collection
The study was conducted in 5 public outpatient primary care centers in Rio de Janeiro, Brazil. Four primary care municipal facilities of Rio de Janeiro City Health Department were included: Copacabana (Centro Municipal de Saúde José de Barros Barreto), Tijuca (Centro Municipal de Saúde Heitor Beltrão), Praça Onze (Centro Municipal de Saúde Marcolindo Candau), and Irajá (Posto Municipal de Saúde Manoel José Ferreira). The Hospital Escola São Francisco de Assis (HESFA), a teaching hospital that is affiliated with Universidade Federal do Rio de Janeiro, was also included. The study was approved by the institutional review boards of the Universidade Federal do Rio de Janeiro, the Rio de Janeiro City Health Department, and the University of Pittsburgh.
Subjects were HIV-1 seropositive adults who had been taking ART for 6 to 24 months. Exclusion criteria were pregnancy, inability to understand or cooperate with protocol requirements, ongoing psychiatric disorders, and apparent intoxication from illicit drugs and/or alcohol at the time of the enrollment visit.
Patients were approached about study participation while awaiting their appointments in the waiting room of the participating public clinic. Following informed consent, subjects underwent an interview using a structured questionnaire that included information about demographic, personal, sexual behavior, and social variables. Factors commonly reported to be associated with ART failure, including measures of self-reported adherence, were also included in the questionnaire. One study investigator (C.B.H.), who was blinded to the patient clinical data at the time, conducted all the interviews.
Following the interview, the patient’s medical record was reviewed. Data abstracted from the medical records included details about prescribed ART medications, clinical information, CD4+ cell counts, and plasma HIV RNA levels. CD4+ lymphocyte counts and plasma HIV RNA levels were measured in laboratories that belong to the Brazilian national network of public laboratories. In these laboratories, lymphocyte subset determinations were made using the Becton Dickinson FACScan procedure, whereas HIV-1 plasma RNA was measured using the NucliSens (BioMérieux SA, Mercyl’Etoile, France/Organon Teknika); the threshold for viral load detection was 80 copies/mL.
Immunologic failure was defined as an increase of less than 0.05 × 109 cells/mL 6 months after the baseline visit. Virologic failure was defined as a reduction of less than 1 log10 HIV-RNA 6 months after the baseline visit. Patients were classified as complete responders (CRs; neither virologic nor immunologic failure), immunologic responders (IRs; without immunologic failure, but with virologic failure), virologic responders (VRs; without virologic failure, but with immunologic failure), and nonresponders (NRs; with both virologic and immunologic failure). Our primary analyses were comparisons of NRs with responders (CR, IR, or VR). These outcomes were selected because it has been shown that patients with VR or NR had a significantly higher risk of progression to AIDS after 24 months of ART when compared with patients with CR. 4,5
Data analysis was performed using Stata version 6.0 statistical software (Stata Corporation, College Station, TX). Univariate analysis was performed using the Wilcoxon (Mann–Whitney) 2-sample test. The Fisher exact test was used to evaluate associations for categoric variables. Variables with a P value ≤ 0.15 were included in the multivariate analysis.
A logistic regression model was fit using the stepwise maximum likelihood estimation technique. The level of significance for removal of a variable in backward regression was 0.10, and forward regression was 0.05. Interactions were assessed using the –2log likelihood ratio test to compare models with and without interactions. The Pearson χ2 goodness of fit test, as well as the Hosmer–Lemeshow test, was used to evaluate fitness of the model.
During a 6-month period, 226 patients were approached about participation and, following written informed consent, 220 (97%) were enrolled. The 6 patients who were not enrolled included 5 who did not wish to participate and 1 who was overtly intoxicated at the time of the interview. An additional 9 enrolled patients (4%) were excluded from further analysis because they did not have available viral load and/or CD4+ cell count measurements after starting ART, leaving 211 patients for analysis.
Of the 211 patients included in the analysis, 67 (32%) were from municipal primary care facilities and 144 (68%) were being followed at HESFA (Table 1). The median age was 37 years, 68% were male, 62% had 8 years or less of formal education, and 54% described their sexual orientation as heterosexual. The only significant difference between the patients from HESFA and the patients from the municipal units was that a higher proportion of patients seen at HESFA were male.
The median baseline CD4+ was 211 cells/mL and the median baseline viral load (log10) was 4.9. The median time on treatment was 16 months. Ninety-two patients (44%) never had an opportunistic disease (OD), 95 patients (45%) experienced at least one OD before starting ART, 10 patients (5%) had an OD after starting ART, and 14 patients (7%) had an OD before and after starting ART. The mean number of pills that should have been taken that was reported as having been taking in the 3 days before the interview was 81% (data not shown).
The 211 patients were classified as follows with respected to response to ART: 173 (82%) were responders and 38 (18%) were NRs. Of the responders, 28 (13% of total) were IRs, 32 (15% of total) were VRs, and 113 (53% of total) were CRs. Ninety-three (44%) reached undetectable viral load. In the univariate analysis, multiple factors were associated with nonresponse to therapy (Table 2).
The multivariate analysis is shown in Table 3. A statistically significant interaction term between baseline CD4+ cell count per milliliter and patient belief that ART would delay HIV progression was included in the model (data not shown). Factors associated with nonresponse to therapy were reported adherence of < 80% in the previous 3 days (odds ratio [OR], 8.6; 95% confidence interval [CI], 2.9–25.7), time between initiation of ART and first laboratory evaluation (OR, 1.4 for each month; 95% CI, 1.1–1.8), opportunistic disease after initiation of ART (OR, 6.8; 95% CI, 1.4–34.0), inability always to read the ART prescription (OR, 3.9; 95% CI, 1.4–10.9), considering that the treating physician was not knowledgeable about HIV (OR, 4.0; 95% CI, 1.1–15.0), believing that therapy makes the patient sick (OR, 5.6; 95% CI, 1.7–18.8), and not having an HIV-positive friend (OR, 6.1; 95% CI, 1.7–21.8). Believing that ART delays HIV disease progression (OR, 0.001; 95% CI, 0.0–0.2) and baseline CD4+ cell count (OR, 0.5; 95% CI, 0.2–1.1 for each increase in 50 CD4+ cells/mL) were associated with a protective effect against ART failure.
To our knowledge, this is the first study on the outcomes of ART in public clinics in a developing country. The most important finding of this study was that the provision of universal access to ART in Brazil, a middle-income developing country, led to levels of response to therapy that were comparable with what has been reported in the United States and western Europe. For example, in the study by Grabar et al 5 of 2236 patients in Europe, 19% had an immunologic response, 17% had a virologic response, 47% had a total response, and 16% had a complete nonresponse—rates that were similar to what we observed in our study. In a study of 213 patients in Los Angeles by Kitchen et al, 6 43% had a complete response, 14% a virologic response, 13% an immunologic response, and 29% had a complete nonresponse. The higher proportion of patients with complete nonresponse in that study is likely because their definition of immunologic response was more strict than ours.
Less than 80% adherence to therapy in the previous 3 days was the factor that was most strongly associated with non-response to therapy. Although self-reported adherence may not be entirely accurate, this finding is in line with other studies that used medication event monitoring systems to measure adherence prospectively. 7–9
We found a strong association between nonresponse and the development of OD in the first 15 months after starting ART. This is consistent with the study of Grabar et al, 5 in which ART-naive patients who did not respond were more likely to have progression of HIV infection. Our results indicate that the criteria by Grabar et al 5 were clinically meaningful in our population.
Several studies demonstrated that higher CD4+ cell counts and lower viral load at the initiation of therapy were associated with a better clinical outcome. 5,6,10,11 In our univariate analysis, nonresponders actually had initially higher CD4+ cell counts and lower viral load. However, in the multivariate model, which included an interaction term for baseline CD4+ cell count and the belief that ART delays HIV disease progression, higher baseline CD4+ cell count and the belief were both associated with less chance of failing therapy. This apparent contradiction can be attributed to the health belief model, in which patients who are more severely ill (ie, have lower CD4+ cell counts) perceive a higher risk of disease progression if they do not take their medications (if they believe in ART), and thus are more adherent to therapy. Gao et al 12 found similar results in a study of 72 patients in varying stages of HIV infection.
We demonstrated that patient perceptions and understanding about ART were an important determinant of therapeutic failure. For example, patients who believed that ART would make them sick had a higher chance of failure and, conversely, patients who believed that ART would delay HIV progression had a lower chance of failure. Inability to read the prescription, an indicator of health literacy, was also an important variable associated with treatment failure in our study. In a cross-sectional study by Kalichman et al, 13 individuals without functional health literacy were significantly less likely to have an undetectable viral load. In a pilot study to evaluate ART prescribing in a clinic in Rio de Janeiro, Carmody et al 14 demonstrated that prescribing more complex ART regimens was associated with failure to pick up the medications. These findings corroborate the importance of patient education about HIV disease to increase adherence to therapy.
Not having an HIV-positive friend was associated with a worse response to therapy in our population. Although the reason for this association is unknown, possibilities include that having a friend who is HIV positive might be an indicator of increased access to social and informational support about HIV, and that seeing a friend experience the effects of HIV could provide motivation to adhere to therapy. Halkitis et al 15 demonstrated that difficulty in communicating with sex partners about HIV was related to missed doses of ART, and consequently a greater chance of failing therapy.
The data from this study suggest that interventions that increase patient knowledge about HIV and ART before therapy is started may improve adherence and response to therapy. Further research in this population will include the development and evaluation of sustainable interventions to improve these outcomes. If successful, such an approach could have a substantial public health impact by reducing HIV-related morbidity and mortality, and possibly reducing the transmission of drug-resistant HIV. 16
In summary, short-term response rates to ART in patients attending public HIV clinics in Rio de Janeiro were comparable with what has been reported for developed countries. These data suggest that middle-income developing countries can successfully implement universal access to ART if appropriate investments in the infrastructure required to monitor therapy are made. Additional studies will be required to determine the long-term sustainability of the Brazilian ART program and whether simple interventions to improve outcomes are effective.
The authors thank the staff of the Hospital Escola São Francisco de Assis, the staff of the Secretaria Municipal de Saúde do Rio de Janeiro, and the Projeto Praça Onze study team for facilitating the study; Ednaura dos Santos; Joyce Snyder and Donna Gibbons for administrative support; and Dr. John Wilson for biostatistical advice.
1. Piot P. Global AIDS epidemic: time to turn the tide. Science
. 2000;288: 2176–2178.
2. Boletim Epidemiológico—Coordenação Nacional de DST/AIDS. Governo Brasileiro. 2000.
3. Galvao J. Access to antiretroviral drugs in Brazil
. 2002;360: 1862–1865.
4. Hogg RS, Heath KV, Yip B, et al. Improved survival among HIV-infected individuals following initiation of antiretroviral therapy
. 1998; 279:450–454.
5. Grabar S, Le Moing V, Goujard C, et al. Clinical outcome of patients with HIV-1 infection according to immunologic and virologic response after 6 months of highly active antiretroviral therapy
. Ann Intern Med
. 2000;133: 401–410.
6. Kitchen CM, Kitchen SG, Dubin JA, et al. Initial virological and immunologic response to highly active antiretroviral therapy
predicts long-term clinical outcome. Clin Infect Dis
7. Paterson DL, Swindells S, Mohr J, et al. Adherence to protease inhibitor therapy and outcomes in patients with HIV infection. Ann Intern Med
8. McNabb J, Ross JW, Abriola K, et al. Adherence to highly active antiretroviral therapy
predicts virologic outcome at an inner-city human immunodeficiency virus clinic. Clin Infect Dis
9. Press N, Tyndall MW, Wood E, et al. Virologic and immunologic response, clinical progression, and highly active antiretroviral therapy
adherence. J Acquir Immune Defic Syndr
. 2002;31(suppl 3):S112–S117.
10. Lucas GM, Chaisson RE, Moore RD. Highly active antiretroviral therapy
in a large urban clinic: risk factors for virologic failure and adverse drug reactions. Ann Intern Med
11. Mezzaroma I, Carlesimo M, Pinter E, et al. Long-term evaluation of T-cell subsets and T-cell function after HAART in advanced stage HIV-1 disease. AIDS
12. Gao X, Nau DP, Rosenbluth SA, et al. The relationship of disease severity, health beliefs and medication adherence among HIV patients. AIDS Care
13. Kalichman SC, Ramachandran B, Catz S. Adherence to combination antiretroviral therapies in HIV patients of low health literacy. J Gen Intern Med
14. Carmody ER, Diaz T, Starling P, et al. An evaluation of antiretroviral HIV/AIDS treatment in Rio de Janeiro public clinic. Trop Med Int Health
15. Halkitis PN, Parsons JT, Wolitski RJ, et al. Characteristics of HIV antiretroviral treatments, access and adherence in an ethnically diverse sample of men who have sex with men. AIDS Care
16. Hecht FM, Robert GM, Petropoulos CJ, et al. Brief report: sexual transmission of an HIV-1 variant resistant to multiple reverse transcriptase and protease inhibitors. New Engl J Med.