Assessing national antiretroviral treatment (ART) program outcomes is essential to determine whether programs have the desired impact and to secure future funding.1,2 In 2007, Rwanda, a small central African country with about 9.3 million persons,3 had an estimated adult HIV seroprevalence of 2.8% and 150,000 persons who were HIV infected, of whom 68,000 were eligible for ART.4,5 National scale-up of free ART by the Government of Rwanda began in January 2004. By December 31, 2007, over 48,000 individuals had initiated ART in Rwanda.6
National ART program evaluation in resource-limited settings is logistically complicated.2,7 Most monitoring and evaluation systems implemented by Ministries of Health, including the one implemented by the Rwandan government, collect summary service statistics from all HIV care and treatment sites, but do not capture national ART patient longitudinal outcome data, such as retention, weight, and CD4+ cell count change. Thus, most published reports of ART patient outcomes in resource-limited settings have come from select cohorts and centers of excellence that may not reflect national program conditions.8-17 We report on key clinical and immunologic outcomes of a nationally representative sample of patients who initiated ART from January 1, 2004, through December, 31, 2005, in the national HIV care and treatment program in Rwanda.
National ART Guidelines
During 2004 and 2005, adults and children were eligible for ART if they were HIV positive and had a diagnosis (1) of World Health Organization (WHO) clinical stage IV, irrespective of CD4+ cell count, (2) WHO clinical stage III and a CD4+ cell count <350 cells per microliter, or (3) WHO clinical stage I or II and a CD4+ T-lymphocyte count <200 cells per microliter.7,18 Patients were also required to meet adherence eligibility criteria, including the identification of a treatment supporter. First-line ART regimens included 2 nucleoside reverse transcriptase inhibitors and a non-nucleoside reverse transcriptase inhibitor, either in individual or in fixed-dose combination tablets. Single-drug substitutions within the same drug class were recommended for toxicity, pregnancy, and incident tuberculosis. Recommended second-line regimens (switches) contained a protease inhibitor and 2 new nucleoside reverse transcriptase inhibitors not previously used in a first-line regimen.
Standardized medical records (the “dossier vert”) were used by all sites providing HIV care and treatment; slightly modified medical records were used by a small number sites supported by implementing partners; however, essential elements of longitudinal patient monitoring were retained. Standardized national ART registers were not in use at all HIV care and treatment sites during the period of evaluation; however, all sites had, or were able to generate, numeric sequential listings of patients initiating ART at the site. Nationally standardized pharmacy files were also introduced on a rolling basis during this period. At some sites, information regarding the most recent visit was recorded in the pharmacy files.
At the time of ART initiation, routine baseline clinical and laboratory assessment of HIV-infected persons in Rwanda include weight and CD4+ cell count monitoring. National clinical HIV care and treatment guidelines recommend that patients initiating ART have monthly clinical follow-up for the first 6 months, followed by routine clinical and CD4+ evaluations at 6-month intervals. Blood specimens for CD4+ evaluations were generally collected at the routine follow-up visit and subsequently recorded in the medical record once test results were returned to the facility.
Primary outcome measures for this analysis were patient retention, mortality, body weight change, and CD4+ cell count change at 6 and 12 months after ART initiation. For this evaluation, “retention” refers to patients being alive and known to be receiving ART at the same clinic where they initiated ART at the end of a follow-up period, “documented death” refers to a death event that was recorded in the medical record, and lost to follow-up (LTFU) refers to a patient who had not been seen at the clinic for any reason for more than 90 days. Patients who were documented as being dead or LTFU, or having stopped treatment or transferred to another clinic, were considered nonretained. For benchmarking purposes, favorable 6- and 12-month outcomes for weight were a ≥5% and ≥10% increase from baseline and for CD4+ cell count change were a ≥80 cells per microliter and ≥120 cells per microliter increase from baseline, respectively.
Study Design, Study Population, and Sampling
A retrospective cohort analysis was performed on data abstracted from the national medical records of adult patients (≥15 years old), who initiated ART from January 1, 2004, through December 31, 2005. By the end of December 2005, a total of 83 ART clinics were providing HIV care and treatment services and about 17,600 adults had initiated ART.
ART clinics with <30 cumulative ART patients enrolled by December 31, 2005, were excluded (n = 12) resulting in inclusion of 71 sites in the ART clinic-sampling frame. A stratified random sampling plan ensured that every patient initiated on ART in Rwanda at these sites within the 2-year period had an equal probability of inclusion in the sample. To address the differences in cumulative ART patient enrollment among sites, site sample selection was stratified by “ART clinic size.” Approximately 20%, 40%, and 60% of patient records were selected for abstraction at large-sized (>1500 patients), medium-sized (175-1500 patients), and small-sized (30-175 patients) clinics, respectively. Based on logistical considerations, desired power, and estimated clinic effect, a sample size of 30 clinics was selected, including the 1 large clinic, 17 of the 34 medium-sized, and 12 of the 36 small-sized clinics. All provinces and 19 of 30 districts in Rwanda were represented in the clinic sample.
Clinic ART patient registries were used as the sampling frames from which patients were randomly selected. Patients were excluded if they had initiated ART before January 1, 2004, or within 6 months of the date of record abstraction or had transferred from another ART site. A random replacement scheme was used for patients with missing or unavailable records, that is, in instances were a patient file was in use by clinical providers or could not otherwise be located or who met exclusion criteria. Based on an 85% 6-month adult retention benchmark, the total number of adults initiated on ART, an estimated design effect of 1.5, desired precision of ±3% or less, and anticipated missing data, a sample size of 3200 patients was considered adequate.
Data were analyzed using SAS 9.1 (SAS Institute, Cary, NC). We compared baseline demographic and clinical characteristics and primary outcome measures using χ2, log rank, and Wilcoxon tests. Given the equal probability of selection of every patient in the study population, direct sample estimates were considered to be unbiased. Logistic regression taking into account the complex sampling (PROC surveylogistic) calculated crude and adjusted odds ratios (AORs) for the associations between age at ART initiation, sex, first-line ART regimen, baseline weight, baseline CD4+ cell count, and the primary outcomes measures at 6 and 12 months after ART initiation.19 In the multivariate analysis, all these predictors, site, and ART clinic size were controlled for, regardless of significance in unadjusted analyses. Kaplan-Meier survival curves estimated the probability of remaining alive and on ART after 3, 6, and 12 months after ART initiation. Censoring occurred at end of study follow-up either based on the date of data collection or other study event that terminated follow-up, such as, transfer to another clinic.
The protocol was approved by the National AIDS Commission (CNLS) Research Committee and the National Ethics Committee in Rwanda and the Institutional Review Boards of the U.S. Centers for Disease Control and Prevention (Atlanta, GA) and Columbia University Mailman School of Public Health (New York, NY).
A total of 3578 records were abstracted, of which 3194 were used. Ninety-six records were excluded from analysis due to ineligibility or missing date of ART initiation and 288 were excluded due to age younger than 15 years. At ART initiation, the median age of adults was 37 years (range: 15-79 years) and 65% were female (Table 1). About 2% of female patients were documented to have been pregnant while on ART. First-line ART regimens included (1) [stavudine (79%) or zidovudine (21%)] plus, (2) [nevirapine (78%) or efavirenz (22%)], and (3) lamivudine. At initiation of ART (baseline), the median weight was 52 kg [interquartile ratio (IQR) 46-58] for female and 57 kg (IQR 51-62) for male patients. Overall, the baseline median CD4+ cell count was 141 cells per microliter (IQR 80-200). Differences were seen in baseline median CD4+ by sex, with 148 cells per microliter (IQR 85-206) for female patients and 129 cells per microliter (IQR 69-189) for male patients (P < 0.001).
Of 3194 adult patients who initiated ART, 2936 [91.9%; 95% confidence interval (CI): 90.3 to 93.5] and 2777 (85.9%; 95% CI: 83.4 to 88.4) remained alive on ART at their original clinic after 6 and 12 months, respectively (Table 2). No differences by sex were seen in the proportion of patients remaining alive on ART at 6 and 12 months. By 6 months, 3.6% (CI: 2.8% to 4.3%) were dead, 3.1% (CI: 1.7% to 4.4%) were LTFU, 0.19% (CI: 0.06% to 0.31%) had stopped treatment, and 1.3% (CI: 0.80% to 1.8%) had transferred out; by 12 months, 4.6% (CI: 3.6% to 5.7%) were dead, 4.9% (CI: 3.2% to 6.6%) were LTFU, 0.27% (CI: 0.13% to 0.42%) had stopped treatment, and 4.3% (CI: 3.1% to 5.6%) were transferred out.
In unadjusted and multivariate analysis, female patients were more likely to remain alive and on ART at 6 months compared with similar male patients (AOR 1.50; 95% CI 1.14 to 1.98) (Tables 3 and 4). Baseline CD4+ cell counts <50 cells per microliter were associated with lower likelihood of retention at 6 months (AOR 0.22: 95% CI: 0.06 to 0.89) and 12 months (AOR 0.25; 95% CI: 0.08 to 0.77), whereas every 3 kg increase in baseline body weight was associated with increased odds of retention at 6 (AOR 1.14; 95% CI: 1.08 to 1.19) and 12 months (AOR 1.10; 95% CI: 1.06 to 1.14). According to Kaplan-Meier survival curves fit to examine time to end of follow-up and stratified by baseline CD4+ cell count, patients with lower CD4+ cell counts were less likely (log rank, <0.0001) to remain alive on ART at 3, 6, and 12 months after ART initiation compared with patients with higher CD4+ cell counts (Fig. 1).
Overall, 3.55% (95% CI: 2.78 to 4.32) and 4.64% (95% CI: 3.63 to 5.66) of adult patients were documented to have died by 6 and 12 months, respectively. Of 3192 persons who initiated ART, 145 were documented deaths during 3966 person-years of follow-up on ART (mortality rate, 3.7 deaths per 100 person-years on ART) and 160 were determined LTFU (combined mortality rate, 7.7 deaths per 100 person-years on ART, if all LTFUs are assumed to be undocumented deaths).
In multivariate analyses of factors associated with documented death by 6 and 12 months after ART initiation, female sex was associated with lower likelihood of death at 6 months (AOR 0.56; 95% CI: 0.37 to 0.84) after ART initiation, every 3 kg increase in baseline weight was associated with reduced odds of death from all causes at 6 (AOR 0.79; 95% CI: 0.73 to 0.86) and 12 months (AOR 0.84; 95% CI: 0.78 to 0.91) after ART initiation, and for every 50 cells per microliter increase in baseline CD4+ mortality was reduced by 6 months (AOR 0.74; 95% CI: 0.63 to 0.87) after ART initiation. When all patients who were LTFU were considered undocumented deaths, baseline CD4+ counts of <50 cells per microliter were associated with substantially increased likelihood of death by 12 months after ART initiation [AOR 4.68 (1.19-18.36)].
Differences were seen in time to documented death by baseline CD4+ cell counts, as patients with baseline CD4+ cell counts ≤50 cells per microliter had a higher probability of death at 3, 6, and 12 months after ART initiation compared with patients with higher counts (log rank, P < 0.001).
Body Weight Change
Overall, median weight increase for both female (excluding those with documented pregnancy after ART initiation) and male patients was 3 kg (IQR 0-6) at 6 months after ART initiation and median weight increased by an additional 1 kg by 12 months for female patients, but was unchanged at 12 months for male patients. In multivariate analysis, for every 50 cells per microliter increase in baseline CD4+, the odds of an increase in weight at 6 and 12 months decreased by 9% (AOR 0.92; CI: 0.87 to 0.93). Likewise, each 3 kg increase in baseline weight was associated with lower odds of experiencing an increase in weight of ≥5% from baseline at 6 months (AOR 0.81; CI: 0.78 to 0.84) and ≥10% from baseline at 12 months (AOR 0.83; CI: 0.81 to 0.86).
CD4+ Cell Count Change
Median CD4+ cell counts increased by 98 cells per microliter (IQR 42-170) at 6 months and 119 cells per microliter (IQR 55-208) after 12 months on ART, relative to baseline. Female patients had median cell count increase at 6 months of 104 cells per microliter (IQR 44-178) compared with a median increase of 89 cells per microliter (IQR 35-150) for male patients (P = 0.005) and at 12 months of 122 cells per microliter (IQR 55-211) compared with 115 cells per microliter (IQR 54-200) for males (P = 0.35). In multivariate analysis, older age at ART initiation (AOR 0.88, per 10-year increment; 95% CI: 0.80 to 0.97) was associated with achieving a CD4+ cell count increase of ≥80 cells per microliter by 6 months on treatment.
Drug Substitution and Switching
Of 3192 patients who initiated ART, 308 (9.6%; 95% CI: 7.2 to 12.1) were documented to have had 1 or more drug substitutions, 272 (79%) adults had 1, and 36 (21%) had 2 consecutive drug substitutions. During 3966 person-years of follow-up on ART, there were 344 substitution events (substitution rate, 8.7 per 100 person-years on ART). Median time from ART initiation to first drug substitution was 284 days (IQR 108-545). Only 7 (<1%) adults were switched (at least 3 new drugs, including a protease inhibitor) to second-line regimens. In multivariate analysis, pregnant (OR 6.24; 95% CI: 2.88 to 13.50) and nonpregnant (OR 1.98; 95% CI: 1.47 to 2.68) females were more likely to have substitution compared with males; age (OR 1.20, per 10-year increase; 95% CI: 1.04 to 1.38), weight (OR 1.10, per 3 kg increase; 95% CI: 1.06 to 1.14), and CD4+ cell count <50 cells per microliter (OR 3.65; 95% CI: 1.18 to 11.32) at ART initiation were associated with having a substitution, although there were no significant differences between stavudine- and zidovudine-containing or nevirapine- and efavirenz-containing first-line regimens.
Overall, 2884 (90%) and 3012 (94%) of patients had baseline weight and CD4+ cell count data available, respectively. Of adult patients remaining alive and on ART at 6 and 12 months, 1664 (57%) and 1349 (49%), respectively, had follow-up weight data available, whereas 1445 (49%) and 957 (35%), respectively, had follow-up CD4+ cell count data available. Comparisons between the baseline characteristics of patients with and without available follow-up weight and CD4+ cell count data showed that the pattern of missing data was random, except that males were more likely than females to have missing data on 12-month CD4+ cell count outcomes (OR 1.24; P = 0.009).
To our knowledge, this study is the first to capture nationally representative outcome data among patients initiating ART as part of global scale-up efforts. Overall, the evaluation was successful at (1) defining the baseline characteristics of patients who initiated ART in Rwanda between January 1, 2004, and December 31, 2005; (2) describing clinical and immunologic outcomes at 6 and 12 months, for patients with available outcome data; and (3) characterizing the completeness of routine patient monitoring data during rapid scale-up.
ART patient retention is an important indicator of program quality. In Rwanda, 92% and 86% of patients remained alive on ART after 6 and 12 months, respectively. In a recent systematic review of literature on ART patient retention in primarily adult scale-up cohorts in sub-Saharan Africa, mean 6- and 12-month retention was 79% and 75%, respectively.17 However, in that review, the authors defined transfer patients as retained, whereas in our study, transfer patients were not considered retained. The percentage of ART patients who were LTFU is strikingly low, at 3.1% and 4.9% of adult patients at 6- and 12-months, respectively. This LTFU is much lower than the average of 15% (range 0%-44%) LTFU in the first year after initiation in the sub-Saharan African scale-up cohorts reported in the review.
Documented adult mortality on ART in Rwanda was relatively low at 6 (3.6%) and 12 (4.6%) months after ART initiation. Documented mortality from Antiviral Therapy in Lower Income Countries (ART-LINC) cohorts was estimated at 6.4% (95% CI: 5.1 to 7.7) and 2.3% (95% CI: 1.5 to 3.2) at 12 months in low-income programs with active and passive follow-up, respectively. This difference suggests that mortality is underestimated in cohorts without active follow-up. Although we were not able to categorize sites in this manner, even if all patients LTFU were actually deaths, total mortality would be about 6.7% at 6 months and 9.5% at 12 months in our study. The mortality rate was 3.7 per 100 person-years follow-up on ART, substantially lower than 16.1 per 100 person-years observed in a large cohort reported from Zambia.11 Thus, mortality within Rwanda's national ART program during the first 2 years of scale-up seemed lower than mortality reported by centers of excellence in similar settings.9,12 As in other resource-limited settings, male patients were more likely to be older, have CD4+ cell counts <50 at ART initiation, and have died by 6 months compared with female patients, suggesting possible differences in health service utilization or other factors.11,16
Adult immunologic outcomes were excellent for those patients with available follow-up data. Compared with ART-LINC data from centers of excellence, adult ART patients in Rwanda had slightly lower median gains in CD4+ cell counts at 6 months (98 vs. 106 cells/μL). The 12-month median gain in CD4+ cell counts was 119 cells per microliter, somewhat less than reports from centers of excellence in resource-limited settings, but suggesting that, on average, adult patients had CD4+ cell counts >250 cells per microliter by 1 year after ART initiation. Data regarding change in body weight over time are challenging to interpret due to the many variables that can affect weight, including pregnancy, nutritional status, and opportunistic infections, such as tuberculosis. However, for those patients with available follow-up data, the median increase in weight of 3 kg at 6 months suggests a modest clinical response to ART, followed by a leveling off thereafter, although the first year of treatment. More data are needed on long-term immunologic responses to ART in Rwanda and elsewhere.
In our evaluation, ART drug substitution rates were low compared with reports from select cohorts in Africa (8.7 vs. 13.0 substitution events per 100 patient-years in Zambia), and switching was rare.11,20,21 The majority of patients initiated stavudine-containing regimens, and these were not associated with drug substitution compared with zidovudine-containing regimens. More information is needed about drug toxicity and other conditions leading to substitution and possible barriers to appropriate switching.
One of the most striking findings from this evaluation is the high level of missing data, particularly for routine weight and CD4+ cell count monitoring. The causes of the missing data are unknown, but may include inadequate training, noncompliance with national clinical guidelines, understaffing, lack of basic supplies and equipment, delays in return of laboratory results, and poor data recording. These findings underscore the need for routine systematic supervision, including mentoring to promote compliance with clinical guidelines and audits of routine data collection tools to promote data quality and completeness.
Limitations primarily relate to the fact that these analyses were based on routinely collected data with incomplete data for weight and CD4+ cell count outcomes. We were unable to quantify the number of missing or unavailable records, although at most sites, the number was considered relatively low. Data regarding additional factors, which may have differentially affected treatment outcomes, such as, tuberculosis22, cotrimoxazole prophylaxis23-25, and adherence, were not consistently available during the period of the evaluation and were not collected.26-28
Rwanda's national ART program achieved excellent 6- and 12-month retention and immunologic outcomes during the first 2 years of rapid scale-up; however, the completeness of outcome data was suboptimal. Nationally representative sampling of routinely available data can provide important information about program quality and may be useful in resource-limited settings. A nationally standardized system of routine systematic supervision at the clinic level is essential to promote quality of care and data collection.29-31
The authors would like to thank Drs Wafaa El-Sadr and Laura Porter for their contributions to the conception of the study. D.L., F.N., E.K., F.N., Don Hoover, J.H., B.E., A.A., T.E., F.S., A.R., J.M., and A.A. initiated the concept of the study. D.L., F.N., E.K., F.N., J.H., A.A., A.R., and F.S. were involved with data collection. D.L., F.N., Wilson Lo, Don Hoover, A.R., B.E., Denis Nash, and T.E. were involved with data analysis and interpretation. All authors contributed to the development of the article and have seen and approved the final version.
1. Boerma JT, Stanecki KA, Newell M-L, et al. Monitoring of scale-up of antiretroviral therapy programmes: methods to estimate coverage. Bull World Health Organ
2. Bennett S, Boerma JT, Brugha R. Scaling up HIV/AIDS evaluation. Lancet
3. Demographic Projections of National Population Census
. Kigali, Kwanda: National Institute of Statistics of Rwanda, 2005.
4. Ministry of Health. HIV-AIDS in Rwanda: 2008 Epidemic Update
. Kigali, Rwanda: Ministry of Health, 2008.
5. Demographic and Health Survey
. Kigali, Rwanda: National Institue of Statistics of Rwanda, 2005.
6. Ministry of Health. TRAC Annual Report
. Kigali, Rwanda: Ministry of Health, 2005.
7. Patient Monitoring Guidelines for HIV Care and Antiretroviral Treatment, WHO
. Geneva, Switzerland: World Health Organization, 2006.
8. Laurent C, Ngom Gueye NF, Ndour CT, et al. ANRS 1215/1290 Study Group, et al. Long-term benefits of highly active antiretroviral therapy in Senegalese HIV-1-infected adults. J Acquir Immune Defic Syndr
9. Ferradini L, Jeannin A, Pinoges L, et al. Scaling up of highly active antiretroviral therapy in a rural district in Malawi: an effectiveness assessment. Lancet
10. Wools-Kaloustian K, Kimaiyo S, Diero L, et al. Viability and effectiveness of large-scale HIV treatment initiatives in sub-Saharan Africa: experience from western Kenya. AIDS
11. Stringer JSA, Zulu I, Levy J, et al. Rapid scale-up of antiretroviral therapy at primary care sites in Zambia. JAMA
12. Severe P, Leger P, Macarthur C, et al. Antiretroviral therapy in a thousand patients with AIDS in Haiti. N Engl J Med
13. Madec Y, Laureillard D, Pinoges L, et al. Response to highly active antiretroviral therapy among severely immuno-compromised HIV-infected patients in Cambodia. AIDS
14. Coetzee D, Hildebrand K, Boulle A, et al. Outcomes after two year of providing antiretroviral treatment in Kayelitsha, South Africa. AIDS
15. ART-LINC Collaboration and ART Cohort Collaboration groups. Mortality of HIV-infected patients in the first year of antiretroviral therapy: comparison between low-income and high-income countries. Lancet
16. Toure S, Kouadio B, Seyler C, et al. Rapid scaling up of antiretroviral therapy in 10,000 adults in Cote d'Ivoire: outcomes and determinants. AIDS
17. Rosen S, Fox MP, Gill CJ. Patient retention in antiretroviral therapy programs in sub-Saharan Africa: a systematic review. PLoS Med
. 2007;4:e298. doi:10.1371/journal.pmed.0040298.
18. Ministry of Health. Guide de Prise en Charge des Personnes Infectees par le VIH au Rwanda
. Kigali, Rwanda: Ministry of Health, 2003.
20. Boulle A, Orrell C, Kaplan R, et al. Substitutions due to antiretroviral toxicity or contraindication in the first 3 years of antiretroviral therapy in a large South African cohort. Antiv Ther
21. Forna F, Liechty C, Solberg P, et al. Clinical toxicity of highly active antiretroviral therapy in a home-based AIDS care program in rural Uganda. J Acquir Immune Defic Syndr
22. Lawn SD, Myer L, Orrell C, et al. Early mortality among adults accessing a community-based antiretroviral service in South Africa: implications for programme design. AIDS
23. van Oosterhout J, Laufer M, Graham S, et al. A community-based study of the incidence of trimethoprim-sulfamethoxazole-preventable infections in Malawian adults living with HIV. J Acquir Immune Defic Syndr
24. Mermin J, Ekwaru JP, Liechty CA, et al. Effect of co-trimoxazole prophylaxis, antiretroviral therapy, and insecticide-treated bed nets on the frequency of malaria in HIV-1-infected adults in Uganda: a prospective cohort study. Lancet
25. Lowrance D, Makombe S, Harries AD, et al. Lower early mortality rates among patients receiving antiretroviral treatment at clinics providing cotrimoxazole prophylaxis in Malawi. J Acquir Immune Defic Syndr
26. Mills EJ, Nachega JB, Buchan I, et al. Adherence to antiretroviral therapy in sub-Saharan Africa and North America: a meta-analysis. JAMA
27. Au JT, Kayitenkore K, Shutes E, et al. Access to adequate nutrition is a major potential obstacle to antiretroviral adherence among HIV-infected individuals in Rwanda. AIDS
28. Bangsberg DR, Perry S, Charlebois ED, et al. Non-adherence to highly active antiretroviral therapy predicts progression to AIDS. AIDS
29. Lowrance D, Filler S, Makombe S, et al. Assessment of a national monitoring and evaluation for rapid expansion of antiretroviral treatment in Malawi. Trop Med Int Health
30. Libamba E, Makombe S, Mhango E, et al. Supervision, monitoring, and evaluation of nationwide scale-up of antiretroviral therapy in Malawi. Bull World Health Organ
31. Makombe S, Hochgesang M, Jahn A, et al. Assessing the quality of data aggregated by antiretroviral treatment clinics in Malawi. Bull World Health Organ