The introduction of highly active antiretroviral therapy (HAART) in highly industrialized Western countries in 1996 led to a marked decline in morbidity and mortality associated with HIV infection, which was reported in the United States, Canada, Europe, and Australia.1-4 HAART has been slower to be adopted in other non-Western countries, principally because of the expense of the drugs.5 The Brazilian Ministry of Health initiated a program of providing antiretroviral therapy (ART) free of charge to HIV-infected persons in 1991.6 Locally produced antiretroviral drugs have been widely used in Brazil since 1993. Nucleoside reverse transcriptase inhibitor (NRTI) monotherapy and then dual-NRTI therapy were provided until 1996, when HAART combining a protease inhibitor (PI) and dual NRTIs was introduced. Brazil now has a unique position as a developing country with more than 160,000 patients receiving ART.7
In North America, Western Europe, and Australia, HAART was initiated at higher CD4+ T-lymphocyte counts (eg, 500 cells/mm3) according to the antiretroviral use guidelines at that time. In Brazil, HAART was more likely to be initiated when the CD4+ T-lymphocyte count was 200 to 300 cells/mm3.8,9 In addition, dual-nucleoside therapy remained an acceptable option for initiation of ART longer than in the United States.9 Brazil also differs from these Western countries in having a significantly higher incidence of tuberculosis.10 For these reasons, it is possible that the effectiveness of HAART in clinical practice might be less in Brazil than in the United States and other highly industrialized countries.
Studies from Brazil indicate that morbidity and mortality from HIV infection has fallen since the introduction of HAART.11-13 There has not been a direct comparison of clinical response between Brazil and the United States, however. Differences in the incidence of specific AIDS-defining infections such as tuberculosis, initiation of HAART at lower CD4+ lymphocyte counts than initially used in the United States, and longer use of dual-NRTI therapy with possible higher rates of resistance could all contribute to differences in clinical disease progression and outcome between Brazil and the United States. We sought to determine if there have been differences in the clinical response to HAART in the United States and Brazil. This analysis could be performed because of a designed similarity in data collection methods in the Johns Hopkins HIV Clinical Practice Cohort in Baltimore, Maryland, and the Evandro Chagas Clinical Research Institute (IPEC) HIV cohort of the Oswaldo Cruz Foundation (FIOCRUZ) in Rio de Janeiro, Brazil.
Description of the Clinical Cohorts
The Johns Hopkins AIDS Service provides care for a large proportion of HIV-infected patients in Baltimore. An observational, longitudinal, clinical database has been maintained on patients receiving primary HIV care since 1990. In this longitudinal database, data are updated regularly using clinic and inpatient clinical documentation (from the Johns Hopkins AIDS Service and elsewhere), laboratory testing results, and pharmacy records. Prescription of ART (drug, dates of use, and dose) is documented by the medical provider and support staff in the clinical records. Trained abstractors record all this information onto standardized forms for processing. Data on mortality come from the medical records, Maryland State vital statistics records, the National Death Index, and the national Social Security Death Index. Details of the methodology have been previously described.14
The IPEC-FIOCRUZ has provided care to HIV-infected patients in Rio de Janeiro since 1986. An observational, longitudinal, clinical database has been maintained on patients receiving primary HIV care in the HIV/AIDS Clinic since 1998. The data collection process was patterned after the process established at the Johns Hopkins AID Service. In this longitudinal database, data are updated regularly using clinic and inpatient clinical documentation, laboratory testing results, and pharmacy records. Prescription of ART (drug, dates of use, and dose) is documented by the medical provider and support staff in the clinical records. Trained abstractors record all this information onto standardized forms for processing. Data on mortality come from the medical records and national death statistics. Details of the methodology have been previously described.15
For the analysis presented, all patients who were enrolled in these clinical practices were included. There were no exclusions.
Race/ethnicity is based on patient self-report in Baltimore and on provider report in Rio de Janeiro. HIV transmission risk group is defined by self-report in both locations. Brazil has adopted its own AIDS case definition,16 but the 1993 Centers for Disease Control and Prevention (CDC) definition for AIDS-defining illness (ADI) was used in this analysis for US and Brazilian data. ADIs included all those in the 1993 CDC criteria.17 Individuals <17 years of age or with fewer than 30 days of follow-up were excluded from the analysis. HAART was defined as 2 NRTIs plus a PI or a nonnucleoside reverse transcriptase inhibitor (NNRTI) or a PI and an NNRTI in combination. A 3-NRTI regimen was not assessed as HAART, because the use of this combination regimen is no longer considered to be an adequately potent regimen.18
Our analyses combined data from Baltimore and Rio de Janeiro into a single database. All variable formatting was compatible. We analyzed data from those patients who started their initial HAART regimen from 1997 to 2004. Our analyses were designed to assess clinical response to the initial HAART regimen. This assessment included the following:
- Suppression of HIV-1 RNA to <400 copies/mL 12 months after starting HAART. The HIV-1 RNA value closest to 12 months was used, with the last obtained value carried forward. If the patient suppressed to <400 copies/mL before each of these time points, he or she was counted as having suppressed even if he or she subsequently had an HIV-1 RNA measurement >400 copies/mL. HIV-1 RNA levels were assessed using the Roche Amplicor (Roche Molecular Diagnostics, Pleasanton, CA) reverse transcriptase polymerase chain reaction (RT-PCR) assay in Baltimore and by nucleic acid sequence-based amplification (NASBA) in Rio de Janeiro.
- Change in CD4+ T-lymphocyte count from baseline at 12 months. The CD4 cell count obtained closest to the time point was used, with the last obtained value carried forward.
- Development of a new AIDS-defining opportunistic infection as defined by the 1993 CDC criteria. This outcome was assessed for a period 2 years after starting HAART.
- Survival (all cause). This outcome was assessed for a period 2 years after starting HAART.
A comparison of the proportion of patients over time who had viral suppression was conducted by means of the χ2 test. The change in CD4 cell count over time was assessed using the t test. Multivariate analysis of HIV-1 RNA suppression at 12 months was done by logistic regression. Multivariate analysis of the change in CD4 cell count at 12 months was done by linear regression. The associations of multiple variables with viral load suppression to <400 copies/mL and CD4 cell count change were delineated. These variables included the patient being antiretroviral naive, baseline CD4 cell counts and HIV-1 RNA levels, HAART regimen (single PI, boosted PI, or NNRTI), and demographic variables (gender, race, and HIV risk group). Finally, if a patient died or was censored because of loss to follow-up before 12 months, the last value of HIV-1 RNA load or the CD4 cell count was carried forward and used. Loss to follow-up averaged only 2.1% by 12 months and 4.8% at 24 months, with no significant differences by year.
A Kaplan-Meier analysis was performed to assess development of an ADI. The log-rank test was used to test for statistical significance of location (Baltimore vs. Rio de Janeiro), stratified by the CD4+ T-lymphocyte count at initiation of HAART. The strata used were ≤50, 51 to 200, and >200 cells/mm3. Multivariate Cox proportional hazards regression was used to assess the associations of location and other demographic and clinical variables with development of an ADI.
A total of 1368 patients from Baltimore and 1045 patients from Rio de Janeiro who started HAART after January 1, 1997, and before January 1, 2005, were studied. A comparison of the demographic and clinical characteristics of the cohorts is shown in Table 1. There were significant differences in race, gender, and HIV transmission risk group, although there was no difference in the proportion with an AIDS-defining condition or in the baseline CD4+ T-lymphocyte count. There was no difference in prior use of an NRTI; however, use of PIs or NNRTIs was different, with more PI use in Rio de Janeiro and more use of NNRTIs in Baltimore as part of the first HAART regimen. First date of HAART use was significantly earlier in Baltimore.
There was no difference in the percentage of patients who achieved an HIV-1 RNA level <400 copies/mL at 12 months after starting HAART (46.9% in Rio de Janeiro, 50.8% in Baltimore), as shown in Table 2. There was also no difference in the log change in HIV-1 RNA level (−1.65 log in Rio de Janeiro, −1.63 log in Baltimore). The change in CD4+ T-lymphocyte count at 12 months after starting HAART was also similar in Rio de Janeiro and Baltimore (116 cells/mm3 in Rio de Janeiro, 122 cells/mm3 in Baltimore; P = 0.54).
A Kaplan-Meier plot of the development of the first ADI after starting HAART is shown in Figure 1, stratified by location (Baltimore vs. Rio de Janeiro) and by CD4+ T-lymphocyte count (≤50, 51-200, and >200 cells/mm3). Patients from Rio de Janeiro contributed 1917 person-years (mean = 1.83 years per patient) and those from Baltimore contributed 2545 patient-years (mean = 1.86 years per patient) of follow-up after starting HAART. There was no significant difference in the development of an ADI between Baltimore and Rio de Janeiro for any of the 3 CD4+ T-lymphocyte strata. The number and percentage of each specific ADI that occurred is shown in Table 3. The most commonly occurring opportunistic illness in Rio de Janeiro was tuberculosis (27.7% of patients), and the most commonly occurring opportunistic illness in Baltimore was esophageal candidiasis (36.8% of patients).
There was no difference in mortality between Rio de Janeiro and Baltimore. In Rio de Janeiro, 70 (6.7%) of 1045 patients died in the first 2 years after starting HAART, and in Baltimore, 115 (8.4%) of 1368 patients died over the same period (P = 0.14 by log-rank test).
Multivariate Cox proportional hazards regression analysis of the development of an ADI is shown in Table 4. There was no difference between Baltimore and Rio de Janeiro in the development of an ADI (relative hazard = 1.02; 95% confidence interval [CI]: 0.82 to 1.25), adjusting for the demographic and clinical variables in the model.
Our analysis found that clinical progression to an ADI (1993 CDC definition) after initial HAART is similar in patients in Rio de Janeiro and patients in Baltimore within each of the CD4+ T-lymphocyte strata present at the time HAART was initiated. The lack of difference in clinical progression between Rio de Janeiro and Baltimore was not affected after adjusting for several other demographic and clinical factors measured in each of the cohorts. The methods of data collection were designed to be similar in each location, and temporal differences in the time of starting HAART were adjusted for in the analysis.
It is notable that the 2 cohorts were similar in regard to extent of clinical disease progression at first use of HAART. The CD4+ lymphocyte level, HIV-1 RNA level, and proportion of patients with a history of opportunistic illness were similar. This suggests that HIV-infected patients tend to present for HIV care relatively late in their course of disease in both locations. As expected, the date of first HAART use was significantly earlier in the Baltimore cohort. Demographically, the 2 cohorts had greater differences by race and particularly by HIV transmission risk group. These factors, however, did not have a significant effect on the time to clinical disease progression in our multivariate analysis.
One reason for conducting this comparative analysis was a concern that initiation of HAART when the CD4+ T-lymphocyte count was 200 to 300 cells/mm3, as was the situation in Brazil when HAART was introduced, would result in a poorer virologic and immunologic response to HAART and subsequent faster disease progression. This potential concern was not supported by the analyses of each of these clinical outcome parameters. In the United States and other Western countries where HAART was introduced as early as 1995, the goal was to initiate HAART at CD4+ T-lymphocyte levels of 500 cells/mm3 or greater (sometimes called “hit hard-hit early”). Analyses of observational data subsequently suggested that waiting to start HAART at a CD4+ T-lymphocyte count of 350 cells/mm3 or even less did not result in higher rates of long-term clinical disease progression, however.19 Based on these and other observational data, guidelines for starting HAART changed in 2001 to recommend waiting to start HAART at lower CD4+ T-lymphocyte counts. The lack of a difference in clinical disease progression between Rio de Janeiro and Baltimore may reflect the fact that waiting to start HAART at a later CD4+ T-lymphocyte level does not make a difference in clinical disease progression when HAART is otherwise used similarly.
Another reason for conducting this analysis was a possible difference in clinical disease progression attributable to differing opportunistic illnesses in the 2 locations. Tuberculosis is much more common in Rio de Janeiro than in Baltimore, reflecting the pandemic of tuberculosis in Brazil and much of the developing world. In addition, cytomegalovirus, toxoplasmosis, and herpes simplex virus (HSV) were reported more commonly in the Brazil cohort. Most other opportunistic illnesses were comparatively less common in Brazil, possibly further reflecting the competing risk of a particularly common illness.20 Even with this difference in the type of opportunistic illnesses that are prevalent in the population, HAART proved similarly effective, however.
Another potential issue was whether there would be a greater frequency of use of NRTI therapy before the start of a HAART regimen in the Rio de Janeiro cohort compared with the Baltimore cohort, given the early availability of NRTI therapy in both locations but the later availability of HAART in Brazil compared with the United States. As it turns out, the prior use of NRTI therapy was almost identical between the 2 locations, and this was not a factor that had a differential impact on clinical outcomes.
We did not assess death as the primary outcome in this analysis; instead we assessed time to development of an opportunistic illness. There was no significant difference in the percentage of deaths between the 2 locations. Additionally, causes of death may be related to HIV infection, but in the United States, death has been increasingly reported to be attributable to other non-HIV-related causes, such as hepatitis and hepatic failure.21 Injection drug use is associated with hepatitis C transmission, and this could be an important confounding factor for any analysis of mortality in the Baltimore cohort. We believed that an analysis of CDC-based opportunistic illness would be a fairer comparison of clinical disease progression. If a patient died before developing an opportunistic illness, his or her follow-up was censored for the survival analysis. For analysis of change in HIV-1 RNA level and CD4+ lymphocyte count, the last value was carried forward. There was no significant difference between the 2 locations in the proportion of deaths within the first 12 months after starting HAART.
Strengths of this analysis include the similarity of procedures and methods to collect and code the data in these clinical cohorts. The Rio de Janeiro cohort was designed to be similar in methods to the Baltimore cohort, making comparison easier and more likely to be homogeneous regarding definitions and criteria for clinical events and completeness of data capture. As clinical cohorts, there were no exclusion criteria and all patients enrolling into these clinics were eligible. As reported in the section on methods, dropout was low in both cohorts. Potential limitations of this analysis include those common to most observational studies. Patients are not randomized to therapy, and it is possible that there is an inherent selection bias as to which patients receive HAART and what specific drugs are prescribed. Rates of HIV-1 RNA suppression are typically lower when reported from clinical practice compared with a clinical trial, probably because of poorer adherence to therapy and other uncontrolled factors. There is a difference in HIV transmission risk groups between the 2 locations, with a relatively high proportion of injection drug users in Baltimore. We did adjust for this factor in our multivariate analysis, without an apparent effect on the similarity in clinical response to HAART between the 2 cities. Additionally, we do not have information on adverse event rates from the antiretroviral regimens received by these patients. Therefore, a comparison of the rate of adverse events could not be done. Finally, we note that the proportion of patients who received NRTI therapy before HAART and the CD4+ lymphocyte count at the start of HAART were not significantly different between the 2 cohorts. We had hypothesized that calendar differences in the use of HAART and differences in CD4+ cell-based guidelines for when to start HAART might have led to differences in response rates. These were not seen, although there were significant differences between the 2 cohorts in the risk of tuberculosis and several other ADIs after HAART was started.
In summary, similar virologic, immunologic, and clinical outcomes were seen in a comparative study of clinical cohorts from Brazil and the United States. We believe this is an optimistic result that may further generalize to other developing countries as HAART is increasingly introduced.
1. Lee ML, Karon MJ, Selik R, et al. Survival after AIDS diagnosis in adolescents and adults during the treatment era, United States 1984-1997. JAMA
2. Hogg RS, Yip B, Chan KJ, et al. Rates of disease progression by baseline CD4 cell count and viral load after initiating triple-drug therapy. JAMA
3. Egger M, May M, Chene G, et al. Prognosis of HIV-1-infected patients starting highly active antiretroviral therapy
: a collaborative analysis of prospective studies. Lancet
4. Dore GJ, Li Y, McDonald A, et al. Impact of highly active antiretroviral therapy
on individual AIDS-defining illness incidence and survival in Australia. J Acquir Immune Defic Syndr
5. Schwartlander B, Stover J, Walker N, et al. Resource needs for HIV/AIDS. Science
6. Dias JLM, Marques R, Leite MG, et al. Antiretroviral therapy
and public health: a balance of the Brazilian experience. In: Access to Antiretroviral Therapy in High Middle Income Latin American Countries
, 1st ed. Licea JAI, ed. Fundacion Mexicana para la Salud, Tepepan, DF, Mexico. 200:75-102. Available at: http://www.sidalac.org.mx/english/publications/arv/contents.htm
. Accessed September 8, 2006.
7. Ministry of Health, Brazil
. Revista Resposta+: Experiência do Programa Brasileiro de AIDS
. Brasília, Brazil
: PN-DST/AIDS/SVS-SVS; 2005.
8. Ministry of Health, Brazil
. Consenso sobre Terapia Anti-retroviral para Adultos e Adolescentes Infectados pelo HIV
. Brazilian Ministry of Health, Brasilia, DF, Brazil
9. Ministry of Health, Brazil
. Recomendações para a Terapia Anti-retroviral em Adultos e Adolescentes Infectados pelo HIV
. Brazilian Ministry of Health, Brasilia, DF, Brazil
10. Fonseca LA, Reingold AL, Casseb JR, et al. AIDS incidence and survival in a hospital-based cohort of asymptomatic HIV seropositive patients in São Paulo, Brazil
. Int J Epidemiol
11. Guerreiro MF, Kerr-Pontes LRS, Mota RS, et al. Survival of adult AIDS patients in a reference hospital of a metropolitan area in Brazil
. Rev Saude Publica
12. Marins JR, Jamal LF, Chen SY, et al. Dramatic improvement in survival among adult Brazilian AIDS patients. AIDS
13. Campos DP, Ribeiro SR, Grinsztejn B, et al. Survival of AIDS patients using two case definitions, Rio de Janeiro, Brazil
, 1986-2003. AIDS
. 2005;19(Suppl 4):S22-S26.
14. Moore RD. Understanding the clinical and economic outcomes of HIV therapy: the Johns Hopkins HIV Clinical Practice Cohort. J Acquir Immune Defic Syndr
15. Campos DP, Veloso VG, Grinsztejn B, et al. Banco de dados de indivíduos HIV positivos para fins de pesquisa clínica: elaboração e atualização. Presented at: 10th Conference on Informatics in Health; 2006; Florianópolis.
16. Ministry of Health, Brazil
. National DST/AIDS Program. AIDS Cases Definition Criteria in Adults and Children
. Manuals series 60. Brazilian Ministry of Health, Brasilia, DF, Brazil
17. US Department of Health Human Services, Centers for Disease Control and Prevention (CDC). Impact of expanded AIDS surveillance case definition on AIDS case reporting. Morb Mort Wkly Rep
18. Gulick R, Ribaudo HJ, Shikuma CM, et al. Triple-nucleoside regimens versus efavirenz-containing regimens for the initial treatment of HIV-1 infection. N Engl J Med
19. Egger M, May M, Chene G, et al. Prognosis of HIV-1-infected patients starting highly active antiretroviral therapy
: a collaborative analysis of prospective studies. Lancet
20. Yan Y, Hoover DR, Moore RD, et al. Competing risk adjustment reduces overestimation of opportunistic infection rates in AIDS. J Clin Epidemiol
21. Palella FJ, Baker RK, Mooreman AC, et al. Mortality in the highly active antiretroviral therapy
era: changing causes of death and disease in the HIV outpatient study. J Acquir Immune Defic Syndr
Keywords:© 2007 Lippincott Williams & Wilkins, Inc.
antiretroviral therapy; Brazil; HIV outcomes; South America