Wester, C William MD*†; Kim, Soyeon MS, DSc Biostat*‡; Bussmann, Hermann MD, MPH*†; Avalos, Ava MD*†; Ndwapi, Ndwapi MD§; Peter, Trevor F PhD, MPH*†; Gaolathe, Tendani MD§; Mujugira, Andrew MD§; Busang, Lesego MS Biostat*∥; Vanderwarker, Chris*†; Cardiello, Peter MD, MPH*†; Johnson, Onalethata PharmD*; Thior, Ibou MD, MS Epid*†; Mazonde, Patson MBChB, MPaed¶; Moffat, Howard FRCP§; Essex, Max PhD, DVM*†; Marlink, Richard MD*†
In developed countries, where HIV-1B predominates, the impressive response to highly active antiretroviral treatment (HAART) is well documented.1-3 Of the 40 million people living with HIV/AIDS worldwide, however, most reside in the developing world and approximately 6 million are in immediate need of antiretroviral (ARV) therapy. In Africa, only 2% of those who need ARV are receiving this potentially life-saving treatment.4,5
Southern Africa has the highest prevalence of HIV/AIDS worldwide,6,7 and is where HIV-1C predominates. HIV-1C has been reported to have higher rates of genomic variation,8 more efficient transcriptional activation,9,10 higher plasma HIV-1 RNA levels,11 and higher rates of mother-to-infant transmission.12 With host, viral, and environmental factors possibly playing an important role, many desire evidence of the effectiveness of HAART in their particular settings. Studies of HAART-treated patients comparing HIV-1-infected patients of African and European descent have shown mixed results in terms of virologic response or disease progression, with some finding poorer outcomes among those of African descent and others finding no significant differences.13-17 Preliminary results among HAART-treated adults are also now available from several African countries, including Senegal, Côte d'Ivoire, Cameroon, Uganda, and South Africa.18-24
In this article, we report on the first adults in Botswana initiating public sector HAART as part of a pilot program.
The Infectious Disease Care Clinic (IDCC) at Princess Marina Hospital (PMH) in Gaborone, Botswana was established to care for HIV-infected adults in the public sector. An ARV treatment program began at the IDCC in April 2001, with patients being referred from the PMH inpatient medical wards, from outpatient medical departments, and from private practitioners. Comprehensive care for HIV-1-infected patients, including prophylaxis for opportunistic infections (OIs), HAART, and laboratory and clinical monitoring, was provided free of charge. All patients with AIDS-defining illnesses and/or CD4+ cell counts <200 cells/mm3 were offered HAART according to the Botswana Guidelines on Antiretroviral Treatment.25 We report on all adult HIV-1-infected ARV-naive patients registered before January 21, 2002, who agreed to receive HAART through this program.
For this program, the HAART regimen used was 2 nucleoside reverse transcriptase inhibitors (NRTIs) plus 1 nonnucleoside reverse transcriptase inhibitor (NNRTI). As per national practice and international guidelines for use of HAART in developing countries at the time, most patients initiated didanosine plus stavudine (ddI + d4T) plus efavirenz (EFV), with nevirapine (NVP) substituted for EFV in women with reproductive potential. Patients initiating treatment after the opening of the National ARV Treatment Program (January 21, 2002) began recommended first-line treatment: Combivir (CBV) plus EFV or NVP.
Before initiation, patients had the following laboratory tests: chemistry, hematology, CD4+ cell count, plasma HIV-1 RNA level, hepatitis B surface antigen, and syphilis serology, with the initial group of consecutively screened patients undergoing additional serologic testing for hepatitis B and C, toxoplasmosis, and cytomegalovirus infection or exposure. Patients were also screened for active OIs, such as tuberculosis (TB) and cytomegalovirus retinitis, if indicated. Patients found to have syphilis, HIV-associated eye disease, or active TB were treated in accordance with national standards.
All patients administered NVP-based HAART were scheduled to return 2 weeks after HAART initiation for review, NVP chemistry blood draw, and dose escalation, unless contraindicated. Blood was drawn for CD4+ cell counts and HIV-1 RNA levels at weeks 4 and 12, and then at 12-week intervals. Blood was drawn for chemistry and hematology at weeks 4, 8, and 12, and then at 12-week intervals.
When patients missed 2 consecutive visits, 2 or more attempts were made to contact them before considering them lost to follow-up.
Patients who qualified for HAART were encouraged to designate a family member or friend as an adherence assistant to aid with ARV medication adherence and toxicity recognition. Once HAART was initiated, patients returned to the IDCC at monthly intervals for medical evaluation, adherence education, and medication refills.
HIV antibody status was determined by use of an enzyme-linked immunosorbent assay (ELISA)-based testing algorithm (Murex HIV 1.2.0 and Ortho Antibody Capture ELISAs performed in parallel). Plasma HIV-1 RNA levels were quantified using the Amplicor HIV-1 Monitor test, version 1.5 (Roche Diagnostics Systems, Branchburg, NJ), with a lower limit of detection of 400 copies/mL. CD4+ cell counts were determined within 4 hours of obtaining the blood sample using the FACSCalibur flow cytometer (Becton Dickinson, San Jose, CA) with CD3/4/8/45 Multiset reagents.
Data were collected from clinical charts and from a hybrid electronic- and paper-based patient management system and were then entered into a separate database for analysis. Clinical data were captured for visits from April 26, 2001 through March 7, 2003. Laboratory data included results of specimens collected through July 18, 2003. Information on loss to follow-up, transfer, and death was available for all patients through November 21, 2003.
An intent-to-treat approach was taken for analyses. Data are presented irrespective of adherence to or changes in ARV regimen. Not all patients, however, were followed for the same length of time because of death, loss to follow-up, transfers, and missing data. Mean changes in CD4+ cell count, weight, and percentage with undetectable HIV-1 RNA level used all available data. Counting time from treatment initiation, nonoverlapping intervals centered at each evaluation time point were created and the measure closest was selected, so that each patient could contribute 1 observation at most to each interval. For each time point, nonoverlapping intervals centered at the time point were created and the measure closest to the targeted time was selected, so that each patient could contribute 1 observation at most to each time point. Graphs show mean changes or percentages with 95% confidence intervals. Because mean changes in CD4+ cell count were well approximated by a normal distribution, 95% confidence intervals used that assumption. For percentages, 95% confidence intervals were calculated using the method of Agresti and Coull.26
Laboratory and clinical toxicities were graded using the Division of AIDS grading scales.27 Grade 2 toxicities were classified as moderate, and grade 3 and 4 toxicities were classified together as severe.
Death, loss to follow-up, and transfer times were censored on November 21, 2003. Time to clinical events and drug switches were censored at the date of the last clinic visit, and time to laboratory marker events were censored at the date of the last laboratory specimen. Kaplan-Meier estimates and 95% confidence intervals based on the Greenwood formula were used to describe time-to-event distributions. The numbers at risk and cumulative numbers of events are given. Log-rank tests and likelihood ratio tests from Cox proportional hazard regression models, which allow for right-censored data, were used to compare time to an event from HAART initiation across groups or for continuous variables and study multiple predictors, respectively.
Using the working definition from a clinical trial that was about to open for enrollment at the site and also because of the fact that Botswana national guidelines did not yet exist, virologic failure was defined as an HIV-1 RNA measurement taken at or after week 12 of >5000 copies/mL. Comparisons by treatment were based on initial regimen and were not censored for regimen changes regardless of reason. All P values are 2-sided and confidence intervals are at the 95% level. No adjustments were made for multiple comparisons.
Ethical approval was obtained from the Botswana Ministry of Health's Health Research Development Committee, PMH Institutional Review Committee, and Harvard School of Public Health's Human Subjects Committee.
Patients Evaluated and Treated
From April 2001 through January 20, 2002, 284 adult patients were seen at the IDCC. Four of the patients were found to be uninfected with HIV, 28 had received prior ARV therapy, and 252 were ARV-naive. Of the 252 HIV-infected ARV-naive patients, 153 were administered HAART, with the rest not initiating HAART for the following reasons: ineligible because CD4+ cell counts were >200 cells/mm3 without an AIDS-defining illness (n = 57), died before receiving HAART (n = 8), lost to follow-up before initiating HAART (n = 33), or refused treatment (1).
Baseline characteristics of the 153 ARV-naive patients are given in Table 1. The median age of the patients was 36 years, and 59% were female. At baseline, most patients had advanced HIV disease, as evidenced by their World Health Organization (WHO) clinical stage, the number of recent or active OIs, and their CD4+ cell count and HIV-1 RNA values.
Fifteen percent had serologic evidence of previous or active syphilis, 9.0% had previous or active hepatitis B exposure, none had hepatitis C exposure, most (94%) had previous cytomegalovirus exposure (positive IgG antibody), and 9% had previous toxoplasmosis exposure (IgG positive for toxoplasmosis antibodies). No patients tested positive for cytomegalovirus or toxoplasmosis IgM antibodies.
Fifty-four (35%) patients had a preexisting diagnosis of peripheral neuropathy, with 98% of the neuropathy classified as mild or moderate (grade 1 or 2).
Antiretroviral Treatment Initiation
The 153 ARV-naive adult patients initiated HAART between April 26, 2001, and December 2, 2002. Most (64%) received ddI + d4T with EFV or NVP.
Follow-Up and Retention
The duration of follow-up was as follows: (1) clinical outcomes, median 54 weeks; (2) laboratory data, median of 63 weeks; and (3) survival, lost to follow up, and patient transfer, median of 96 weeks. The Kaplan-Meier estimate of loss to follow-up at 1 year was 8.4% (3.7%, 12.8%). In this cohort, no patients who remained on follow-up beyond a year were subsequently lost. Some patients transferred to other government ARV treatment clinics, and the 1-year Kaplan-Meier loss to follow-up plus transfer estimate was 13.6% (7.7%, 19.1%).
Response to Treatment
The mean CD4+ cell count increase was 119 cells/mm3 after 4 weeks of HAART and 204 cells/mm3 after 48 weeks, corresponding to a mean absolute CD4+ cell count of 311 cells/mm3 for the group (Fig. 1a).
Four weeks after HAART initiation, 34.6% of patients had undetectable HIV-1 RNA levels. By week 24, 87.0% of patients had undetectable HIV-1 RNA levels. Thereafter, the percentage diminished to 78.8% having undetectable HIV-1 RNA levels at week 48 (see Fig. 1b).
The mean body weight increase at week 12 after HAART initiation was 2.2 kg (1.0, 3.3; n = 68). The mean weight increases at weeks 24 and 48 were 4.4 kg (2.9, 5.9; n = 68) and 4.3 kg (2.7, 5.9; n = 77), respectively. Mean weight increase by week 48 among patients with wasting syndrome at baseline (n = 59) was 7.0 kg (4.5, 9.6), and among patients without wasting syndrome at baseline (n = 84), it was 2.2 kg (0.3, 4.0). Changes in weight were significantly more pronounced among those with wasting syndrome at all time points (P < 0.05; data not shown).
During follow-up, 22 patients experienced virologic failure (HIV-1 RNA value >5000 copies/mL), and the 1-year Kaplan-Meier estimate of the percentage with virologic failure was 15.3% (7.5%, 22.5%).
In 9 of the 22 patients classified with virologic failure, ARV treatment was interrupted for the following reasons: OIs (n = 3), toxicity (n = 5), or both (n = 1). Temporary treatment holds were placed on patients with active OIs, including severe, refractory gastroenteritis and/or wasting syndrome. Toxicity-related treatment holds were primarily for peripheral neuropathy, hepatotoxicity, or pancreatitis. Of the 13 patients with HIV-1 RNA levels >5000 copies/mL without ARV treatment holds, 9 had primary adherence problems, which were classified as follows: 3 because of stigma, 2 living far from the clinic, 1 alcohol related, and 3 unknown. The 4 patients perceived not to have adherence problems were switched to 2 unused NRTIs plus nelfinavir.
Twenty-four of the 153 patients died during follow-up. The Kaplan Meier 1-year survival estimate was 84.7% (79.0%, 90.8%; Fig. 2a). Of those who survived 6 months after HAART initiation, 94.1% (90.0%, 98.5%) survived an additional year.
When comparing survival by baseline CD4+ cell count, we found a 3.2 (1.4, 7.5)-fold higher risk of mortality among patients initiated on ARV treatment with CD4+ cell counts <50 cells/mm3 compared with those with CD4+ cell counts ≥50 cells/mm3 (see Fig. 2b; P = 0.004). The Kaplan Meier 1-year survival estimate was 76.0% (64.5%, 89.4%) for patients with baseline CD4+ counts <50 cells/mm3 versus 90.3% (84.4%, 96.5%) for patients with CD4+ counts ≥50 cells/mm3.
In addition to CD4+ cell count, a higher Karnofsky score (on a continuous scale) was the only other significant univariate predictor (P = 0.016) of better survival. Baseline log HIV-1 RNA level was not a significant predictor (P = 0.093), nor was symptomatic WHO clinical status (P = 0.14) or history of OIs (all P > 0.10). In a Cox regression model that included CD4+ cell count and Karnofsky score, CD4+ cell count remained a statistically significant predictor (P = 0.039) but Karnofsky score did not (P = 0.11).
Autopsy data were not available; however, of the 24 deaths, the most common cause of death was advanced AIDS (n = 4) and pulmonary TB (n = 4). Other patients died of disseminated and/or visceral Kaposi sarcoma (n = 2), hepatotoxicity (n = 2), severe wasting syndrome and/or refractory gastroenteritis (n = 2), cryptococcal meningitis (n = 1), anemia (n = 1), small bowel lymphoma (n = 1), renal failure (n = 1), suicide (n = 1), and toxicity from traditional medicine use (n = 1). For 4 patients (17%), the reason for death was unknown.
Forty-five people had 64 new OIs while on HAART. The 1-year Kaplan-Meier estimate of the percentage experiencing at least 1 new OI was 32.5% (23.4%, 40.6%). Few patients received isoniazid preventative therapy (IPT) for TB prevention, and TB was the most common OI, with 12 patients experiencing a new diagnosis. The Kaplan-Meier estimate of new TB diagnoses at 1 year was 8.8% (4.3%, 15.0%). When analyzed by baseline CD4+ cell count, there was no significant difference in rates of TB diagnosis (CD4+ count <50 vs. ≥50 cells/mm3; P = 0.62). Other OIs diagnosed after HAART initiation included herpes zoster (n = 8), oral and/or esophageal candidiasis (n = 8), wasting syndrome (n = 6), and diarrhea lasting 7 or more days (n = 6). Lower CD4+ cell count and low weight (<40 kg) were associated with increased incidence of oral or esophageal candidiasis, and low weight was associated with increased incidence of herpes zoster.
Eighteen people experienced 23 OIs within the first 4 months of HAART initiation. Of these OIs, 13 of the 23 were attributed to immune reconstitution,24 including 5 cases of pulmonary TB, 4 cases of cutaneous herpes zoster, 2 cases of Pneumocystis jiroveci pneumonia (PCP), 1 case of cryptococcal meningitis, and 1 case of unspecified bacterial pneumonia.
The following toxicities of moderate or worse severity, listed in decreasing order of frequency, were most frequently seen: peripheral neuropathy, skin rash and/or hypersensitivity reaction, pancreatitis, hepatotoxicity, and anemia.
Thirty-four of the 153 patients had a least 1 severe toxicity, and the Kaplan-Meier estimate of the percentage experiencing severe toxicities by 1 year was 23.8% (15.5%, 31.2%; Fig. 3a).
Two patients on zidovudine plus lamivudine (ZDV + 3TC)-containing HAART developed severe anemia, necessitating a treatment switch after hospitalization for blood transfusion.
Evaluating peripheral neuropathy of moderate or severe grade showed an interaction between a preexisting diagnosis of peripheral neuropathy and whether or not the patient was placed on a ddI + d4T-containing regimen (P = 0.024; see Fig. 3b). Patients without a preexisting diagnosis of peripheral neuropathy placed on non-ddI + d4T-containing regimens had lower rates of moderate or severe peripheral neuropathy compared with those with a preexisting diagnosis of peripheral neuropathy and/or those placed on ddI + d4T-containing regimens. Patients without preexisting peripheral neuropathy had a 4.3 (1.7, 11.5)-fold higher rate of developing moderate or severe peripheral neuropathy if placed on ddI + d4T-containing regimens. This was not the case for patients placed on ddI + d4T-containing regimens with a preexisting diagnosis of peripheral neuropathy (P = 0.92). Among those treated with non-ddI + d4T-containing regimens, preexisting peripheral neuropathy was associated with a 4.4 (1.1, 18.0)-fold higher risk of developing moderate or severe peripheral neuropathy (P = 0.034). Multivariable Cox models showed that CD4+ cell count, viral load, WHO clinical stage, and weight did not explain any differences.
Patients experienced severe hepatotoxicity as defined by >5-fold elevations in serum glutamic-oxaloacetic transaminase (SGOT)/aspartase aminotransferase (AST) or severe serum glutamic-pyruvic transaminase (SGPT)/alanine aminotransferase (ALT) levels at an estimated rate of 4.5 (2.2, 9.5) toxicities per 100 person-years and 3.2 (1.3, 7.7) toxicities per 100 person-years, respectively.
Forty-seven of the 153 patients had 49 toxicities that led to an ARV drug switch (2 patients had 2 switches), and the Kaplan-Meier estimate of the percentage of patients with toxicities leading to drug switches by 1 year was 32.2% (23.0%, 40.4%). Of the 49 drug switches, 14 (29%) were for severe peripheral neuropathy, with all patients on ddI + d4T-containing HAART; 17 (35%) were for moderate peripheral neuropathy, with most patients on ddI + d4T-containing HAART; 3 (6%) were for hepatoxicity, with 2 patients on NVP-containing HAART; 2 (4%) were for pancreatitis, with both patients on ddI-containing HAART; 2 (4%) were for NVP-induced severe hypersensitivity and 5 (11%) were for NVP-induced moderate rashes; 2 (4%) were for severe anemia; and 4 (8%) were for other or unknown reasons. Of these 49 drug switches, 29 were for severe toxicities.
Toxicity-driven drug switch rates did not differ by baseline CD4+ cell count <50 or ≥50 cells/mm3 (P = 0.35) and was significantly higher for ddI + d4T-containing regimens but not according to whether NVP was contained in the regimen (P = 0.28).
The initial virologic response to protease-sparing HAART regimens among treatment-naive adults in Botswana known from other studies to be primarily infected with HIV-1 subtype C28-30 was similar to that documented in other countries representing diverse viral subtypes, host genetic profiles, socioeconomic groups, and care delivery mechanisms.13,20-23,31-35 At 48 weeks, the mean CD4+ increase was 204 cells/mm3 and 78.8% of patients had undetectable plasma HIV-1 RNA levels.
Twenty-two patients had HIV-1 RNA levels >5000 copies/mL. After accounting for drug holds and adherence problems, only 4 patients were considered to have had virologic failure as a possible result of drug resistance. Most of those with HIV-1 RNA values >5000 copies/mL actually had adherence problems, drug toxicities, or active OIs necessitating treatment holds. Among those identified with adherence problems, interviews suggested that the primary causes included stigma and social issues, geographic constraints, and alcohol abuse.
Mortality was 3.2-fold higher among those initiated with baseline CD4+ cell counts <50 cells/mm3, which is consistent with the existing literature.36-38 CD4+ cell count was the only independent predictor of survival, although the small sample size may have limited our ability to detect smaller effects. Whenever possible, HIV-1-infected adults should be referred before their CD4+ cell count reaches levels <50 cells/mm3, which should significantly improve their chances of survival. Despite their poorer prognosis, 76% of patients with CD4+ cell counts <50 cells/mm3 survived at least 1 year.
Most patients had many preexisting or recent OIs before treatment initiation, and many experienced new OIs, approximately 50% of which were attributed to immune reconstitution. The spectrum of baseline OIs differed from HIV-1B-infected patients in that few patients had PCP but many had TB, Kaposi's sarcoma, and wasting syndrome, which was usually not associated with chronic diarrhea, contrary to other African reports.39,40 Since the program began, Botswana has scaled up its IPT program, offering 6 months of isoniazid to all qualifying HIV-1-infected adults before their referral for HAART initiation.
Patients in our program had high rates of treatment-modifying toxicities, with most occurring within the first year on treatment. Of the 57 patients who experienced a new diagnosis of peripheral neuropathy, 20 had a preexisting diagnosis of peripheral neuropathy. Peripheral neuropathy was commonly seen among those taking ddI + d4T and among those with a preexisting diagnosis of peripheral neuropathy. Among those without a preexisting diagnosis of peripheral neuropathy, there was a 4.3-fold increase in the rate of moderate or severe peripheral neuropathy. Our rate of treatment-modifying peripheral neuropathy was similar to that in patients studied in AIDS Clinical Trials Group Study 384,31 suggesting that adults in Botswana experienced similar rates of this complication. Although ddI + d4T is no longer recommended as a dual-NRTI backbone of first-line HAART, this combination is still used in second-line HAART regimens. The significant risk of debilitating peripheral neuropathy as well as the enhanced potential for cross-resistance between the thymidine analogues d4T and ZDV among patients who fail the first-line National ARV Treatment Program regimen (ZDV + 3TC plus EFV or NVP) has significant implications for the choice of second-line treatments, which might include ddI + d4T, especially among those who have a preexisting diagnosis of peripheral neuropathy.
Analysis data were checked for consistency and compared with charts for resolution of discrepancies. All deaths were reviewed, and patients were confirmed for possible death before designating them lost to follow-up. Because some data were collected from charts rather than prospectively collected and managed, it is possible that we underreported toxicities and OIs. For patients who died, it is possible that OIs and toxicity are also underreported, because not all were hospitalized before death and we often relied on information obtained from family members. Additionally missing data could have biased the results, particularly if more complex patients had missing charts. We also relied on the existing patient management system for up-to-date survival, loss to follow-up, and transfer information. CD4+ cell counts and plasma HIV-1 RNA measures were sometimes not monitored according to recommended schedules or were missing from the charts.
We compared rates of toxicities by whether or not ddI + d4T was used and whether or not NVP was used, and, when possible, we tested for interactions. Because patients were not randomized to treatment received, results must be interpreted with caution. Patients receiving regimens containing CBV tended to be enrolled later, and those with preexisting peripheral neuropathy tended not to be initiated on ddI + d4T-containing regimens. Although attempts to correct for imbalances in baseline factors were made using models, not all factors were known or adequately adjusted for.
We did not systematically collect data on adherence; therefore, we cannot provide data on overall adherence rates. Only when patients had HIV-1 RNA levels >5000 copies/mL did we extensively review contributing factors.
The successful outcomes of adults treated in our pilot program are significant because Botswana has the first operational National ARV Treatment Program in Africa. As of August 1, 2004, our urban ARV treatment clinic's capacity has grown exponentially, now providing care to more than 13,000 adults, more than 10,000 of whom receive HAART. Interestingly, the recently published results in Khayelitsha20 show a high degree of concordance with ours, despite the fact that patients in their study were highly prescreened and thus more likely to have excellent adherence rates. The success of the IDCC and of programs in Africa13,20-23,33-35,41 suggests that it is possible to open large-scale public outpatient ARV treatment clinics in other developing countries. Nevertheless, it must be acknowledged that our ARV treatment clinic was established at a large urban referral hospital in an environment with a strong health delivery system. There are significant logistical and operational issues involved in the establishment of a large ARV treatment clinic, some of which are discussed for the Botswana context in another article.42 Still, these results should provide valuable information to those who are responsible for scaling up the provision of HAART through various programs, including the WHO's 3 × 5 initiative.5,43,44
The authors thank the patients who participated in this initial public sector ARV program. They are grateful to Dr. Diana Dickinson and her staff from Independence Surgery for providing pharmaceutical support while the critical pharmacy infrastructure was being developed; to the Tshepo Study Staff of the Botswana-Harvard School of Public Health AIDS Initiative Partnership, who provided key clinical and counseling staff; to the entire IDCC staff; and to the seconded physicians from the WHO and University of Pennsylvania via the African Comprehensive HIV/AIDS Partnership, who participated in patient care and teaching. The authors thank Joy Phumaphi, former Minister of Health, and the entire Ministry of Health of Botswana for supporting this program, and Bristol-Myers Squibb's Secure the Future foundation for creating the initial laboratory and clinical infrastructure. They also acknowledge and thank Patrick Finch, Katie Holland (Harvard School of Public Health, Boston, MA) for their editorial assistance. Ava Avalos and Ibou Thior are former Fellows of the Fogarty International Center grant TW00004.
1. Palella FJ, Jr, Delaney KM, Moorman AC, et al. Declining morbidity and mortality among patients with advanced human immunodeficiency infection. N Engl J Med. 1998;338:1933-1942.
2. Mocroft A, Vella S, Benfield TL, et al. Changing patterns of mortality across Europe in patients with human immunodeficiency infection. Lancet. 1998;352:1725-1730.
3. Moore RD, Chaisson RE. Natural history of HIV infection in the era of combination antiretroviral therapy. AIDS. 1999;13:1933-1942.
4. UNAIDS. Fact sheet. Access to HIV treatment and care, 2003. Available at: www.unaids.org
5. The WHO and UNAIDS Global Initiative to provide antiretroviral therapy to 3 million people with HIV/AIDS in developing countries by the end of 2005. WHO/UNAIDS Report: treating 3 million by 2005. Making it happen, the WHO strategy. 2003. Available at: http://www.unaids.org
7. Essex M, M'Boup S. Regional variations in the African epidemics. In: Essex M, M'Boup S, Kanki PJ, et al, eds. AIDS in Africa. 2nd ed. New York: Kluwer; 2002:631-640.
8. Novitsky VA, Montano MA, McLane MF, et al. Molecular cloning and phylogenetic analysis of human immunodeficiency virus type 1 subtype C: a set of 23 full-length clones from Botswana. J Virol. 1999;73:4427-4432.
9. Montano MA, Novitsky VA, Blackard JT, et al. Divergent transcriptional regulation among expanding human immunodeficiency virus type 1 subtypes. J Virol. 1997;71:8657-8665.
10. Montano MA, Nixon CP, Ndung'u T, et al. Elevated tumor necrosis factor-alpha activation of human immunodeficiency virus type 1 subtype C in Southern Africa is associated with an NF-kappaB enhancer gain-of-function. J Infect Dis. 2000;181:76-81.
11. Neilson JR, John GC, Carr JK, et al. Subtypes of human immunodeficiency virus type 1 and disease stage among women in Nairobi, Kenya. J Virol. 1999;73:4393-4403.
12. Blackard JT, Renjifo B, Fawzi W, et al. HIV-1 LTR subtype and perinatal transmission. Virology. 2001;287:261-265.
13. Nellen JF, Wit FW, de Wolf F, et al. Virologic and immunologic response to highly active antiretroviral therapy in indigenous and nonindigenous HIV-1-infected patients in The Netherlands. J Acquir Immune Defic Syndr. 2004;36:943-950.
14. Frater AJ, Dunn DT, Beardall AJ, et al. Comparative response of African HIV-1 infected individuals to highly active antiretroviral therapy. AIDS. 2002;16:1139-1146.
15. Staehelin C, Rickenbach M, Low N, et al. Migrants from sub-Saharan Africa in the Swiss HIV Cohort Study: access to antiretroviral therapy, disease progression and survival. AIDS. 2003;17:2237-2244.
16. Morgan D, Mahe C, Mayanja B, et al. Progression to symptomatic disease in people infected with HIV-1 in rural Uganda: prospective cohort study. BMJ. 2002;324:193-196.
17. del Amo J, Petruckevitch A, Phillips A, et al. Disease progression and survival in HIV-1 infected Africans in London. AIDS. 1998;12(10):1203-1209.
18. Gaston D, Roels T, Ellerbrock T, et al. Virologic and immunologic outcomes and programmatic challenges of an antiretroviral treatment pilot project in Abidjan, Côte d'Ivoire. AIDS. 2003;17(Suppl 3):5-15.
19. Laniece I, Ciss M, Desclaux A, et al. Adherence to HAART and its principal determinants in a cohort of Senegalese adults. AIDS. 2003;17(Suppl 3):103-108.
20. Coetzee D, Hildebrand K, Boulle A, et al. Outcomes after two years of providing antiretroviral treatment in Khayelitsha, South Africa. AIDS. 2004;18:887-895.
21. 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. 2002;360:34-40.
22. Laurent C, Diakhate N, Gueye NF, et al. The Senegalese government's highly active antiretroviral therapy initiative: an 18-month follow-up study. AIDS. 2002;16:1363-1370.
23. Djomand G, Roels T, Ellerbrok T, et al. Virologic and immunologic outcomes and programmatic challenges of an antiretroviral treatment pilot project in Abidjan, Côte d'Ivoire. AIDS. 2003;18(Suppl 3):5-15.
24. Landman R, Schiemann R, Thiam S, et al. Once-a-day highly active antiretroviral therapy in treatment-naive HIV-1-infected adults in Senegal. AIDS. 2003;17:1017-1022.
25. Anabwani G, Jimbo W, eds. Botswana Guidelines on Antiretroviral Treatment. Gaborone: Botswana Ministry of Health; 2002.
26. Agresti A, Coull BA. Approximate is better than exact for interval estimation of binomial proportions. Am Stat. 1998;52:119-126.
27. Division of AIDS, NIAID. Division of AIDS Table for Grading Severity of Adult Adverse Experiences. Rockville, MD: National Institute of Allergy and Infectious Diseases; 1992.
28. Novitsky VA, Montano MA, McLane MF, et al. Molecular cloning and phylogenetic analysis of HIV-1 subtype C: a set of 23 full-length clones from Botswana. J Virol. 1999;73:4427-4432.
29. Ndung'u T, Renjifo B, Novitsky VA, et al. Molecular cloning and biological characterisation of full-length HIV-1C subtype C from Botswana. Virology. 2000;278:390-399.
30. Doualla-Bell F, Gaseitsiwe S, Ndung'u T, et al. Mutations and polymorphisms associated with antiretroviral drugs in HIV-1C-infected African patients. Antivir Chem Chemother. 2004;15:189-200.
31. Hirsch HH, Kaufmann G, Sendi P, et al. Immune reconstitution in HIV-infected patients. Clin Infect Dis. 2004;38:1159-1166.
32. Shafer RW, Smeaton LM, Robbins GK, et al. Comparison of four-drug regimens and pairs of sequential three-drug regimens as initial therapy for HIV infection. N Engl J Med. 2003;394:2304-2315.
33. Vinh-Kim N, Grennan T, Peschard K, et al. Antiretroviral use in Ouagadougou, Burkina Faso. AIDS. 2003;17(Suppl 3):109-111.
34. Staszewski S, Morales-Ramirez J, Tashima KT, et al. Efavirenz plus zidovudine and lamivudine, efavirenz plus indinavir, and indinavir plus zidovudine and lamivudine in the treatment of HIV-1 infection in adults. N Engl J Med. 2003;341:1865-1873.
35. Montaner JSG, Reiss P, Cooper D, et al. A randomized, double-blind trial comparing combinations of nevirapine, didanosine, and zidovudine for HIV-infected patients; the INCAS Trial. JAMA. 1998;279:930-937.
36. Mujugira A, Wester CW, Kim S, et al. Antiretroviral treatment among ARV naïve HIV-1 subtype C infected adults with CD4 <50 cells/mm3 at treatment initiation [poster presentation]. Presented at: XV International AIDS Conference; 2004; Bangkok.
37. Miller V, Phillips AN, Clotet B, et al. Association of virus load, CD4 cell count, and treatment with clinical progression in human immunodeficiency virus-infected patients with very low CD4 cell counts. J Infect Dis. 2002;186:189-197.
38. Sabin CA, Smith CJ, Gumley H, et al. Late presenters in the era of highly active antiretroviral therapy: uptake of and responses to antiretroviral therapy. AIDS. 2004;18:2145-2151.
39. Ledru E, Sanou PT, Dembele M, et al. Prevention of wasting and opportunistic infections in HIV-infected patients in West Africa: a realistic and necessary strategy before antiretroviral treatment. Sante. 1999;9:293-300.
40. Lucas SB, DeCock KM, Hounnou A, et al. Contribution of tuberculosis to slim disease in Africa. BMJ. 1994;308:1531-1533.
41. Freidl AC, Lederberger B, Flepp M, et al. Response to first protease-inhibitor- and efavirenz-containing antiretroviral combination therapy; the Swiss HIV Cohort Study. AIDS. 2001;15:1793-1800.
42. Wester CW, Bussmann H, Avalos A, et al. Establishment of a public adult antiretroviral treatment clinic in urban Botswana: lessons learned. J Clin Infect Dis. 2005;40(6).
43. Global HIV/AIDS Program of the World Bank. Provision of ARV therapy in resource limited settings: the challenges of drug resistance and adherence. June 17-18, 2003.
44. Joint United Nations Programme on HIV/AIDS (UNAIDS). UNAIDS report: accelerating action against AIDS in Africa, 2003. Available at: http://www.unaids.org
© 2005 Lippincott Williams & Wilkins, Inc.