Objectives: To measure rates and predictors of virologic failure and switch to second-line antiretroviral therapy (ART) in South Africa.
Design: Observational cohort study.
Methods: We included ART-naive adult patients initiated on public sector ART (January 2000 to July 2008) at 5 sites in South Africa who completed ≥6 months of follow-up. We estimated cumulative risk of virologic failure (viral load ≥400 copies/mL with confirmation above varying thresholds) and switching to second-line ART.
Results: Nineteen thousand six hundred forty-five patients (29,935 person-years) had a median of 1.3 years of study follow-up (1.8 years on ART) and a median CD4 count of 93 (IQR: 39–155) cells per microliter at ART initiation. About 9.9% (4.5 per 100 person-years) failed ART in median 16 (IQR: 12–23) months since ART initiation, with median 2.7 months (IQR: 1.6–4.7) months between first elevated and confirmatory viral loads. By survival analysis, using a confirmatory threshold of 400 copies per milliliter, 16.9% [95% confidence interval (CI): 15.4% to 18.6%] failed by 5 years on ART, but only 7.8% (95% CI: 6.6% to 9.3%) using a threshold of 10,000. CD4 <25 versus 100–199 (adjusted HR: 1.60; 95% CI: 1.37 to 1.87), ART initiation viral load ≥1,000,000 versus <10,000, (1.32; 0.91 to 1.93), and 2+ gaps in care versus 0 (95% CI: 7.25; 4.95 to 10.6) were predictive of failure. Overall, 10.1% (95% CI: 9.0% to 11.4%) switched to second-line by 5 years on ART. Lower CD4 at failure and higher rate of CD4 decline were predictive of switch (decline 100% to 51% versus 25% to –25%, adjusted HR: 1.96; 95% CI: 1.35 to 2.85).
Conclusions: In resource-limited settings with viral load monitoring, virologic failure rates are highly sensitive to thresholds for confirmation. Despite clear guidelines there is considerable variability in switching failing patients, partially in response to immunologic status and postfailure evolution.
*Center for Global Health and Development, Boston University, Boston, MA
†Department of Epidemiology, Boston University School of Public Health, Boston University, Boston, MA
‡Medicines Sans Frontier, Cape Town, South Africa
§McCord Hospital, Durban, South Africa
‖Clinical HIV Research Unit, University of the Witwatersrand, Johannesburg, South Africa
¶Institute of Social and Preventive Medicine, University of Bern, Bern, Switzerland
#Department of Medicine, University of Stellenbosch and Tygerberg Academic Hospital, Cape Town, South Africa
**Desmond Tutu HIV Centre, University of Cape Town, Cape Town, South Africa
††Department of Epidemiology, Harvard School of Public Health, Boston, MA
‡‡Harvard-MIT Division of Health Sciences and Technology, Boston, MA
§§Department of Social Medicine, School of Social and Community Medicine, University of Bristol, Bristol, United Kingdom
‖‖Centre for Infectious Disease Epidemiology and Research, School of Public Health and Family Medicine, University of Cape Town, Cape Town, South Africa.
Correspondence to: Matthew P. Fox, DSc, MPH, Crosstown Center, 3rd Floor, 801 Massachusetts Ave, Boston, MA 02118 (e-mail: email@example.com).
Supported by Grant Number U01AI069924 from NIH (NIAID, NICHD, NCI, Principal Investigators Egger and Davies). Dr Fox was funded by Award Number K01AI083097 from the National Institute of Allergy And Infectious Diseases. Dr Hernán was funded by NIH R01 AI073127. Professor Sterne was supported by UK MRC grant G0700820. Dr Keiser was funded by a PROSPER fellowship by the Swiss National Science Foundation (Grant 32333B_131629). The findings are solely the responsibility of the authors and do not necessarily represent the official views of the NIH. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the article.
The authors have no conflicts of interest to disclose.
The Members of the IeDEA Southern Africa steering group and site principal investigators are listed in Appendix 3.
Received November 28, 2011
Accepted February 17, 2012
As the global scale-up of antiretroviral therapy (ART) has reached nearly 5 million people,1 a growing body of evidence from large observational cohorts has demonstrated positive clinical, immunologic, and virologic outcomes being achieved throughout sub-Saharan Africa.2–7 Even though large numbers of new patients are still starting treatment in resource-limited settings, focus is shifting from the short-term stresses of treatment initiation to the long-term problems of managing a lifelong chronic disease. A critical part of this shift is an emphasis on managing the growing number of public-sector patients who already have failed or will soon fail first-line therapy.8–13
Although recent estimates suggest only 2% of those currently on ART are on a second-line regimen,14 a far greater number are likely to be failing virologically but have not switched from first-line therapy: WHO estimates that 500,000 to 800,000 patients required switching to second-line regimens by 2010.15 As ART scale-up continues and the average duration on ART increases, both the absolute number and relative proportion of patients needing second-line therapy continues to grow. It has previously been estimated in South Africa that by 5 years on ART, 14% of patients fail virologically.2
Poor access to HIV-1 RNA viral load testing is a key challenge facing many national programs12,16 and therefore, reliance on clinical and immunologic means of determining when to switch, which have poor predictive ability for virologic failure, is problematic.17–22 There is increasing pressure to improve access to viral load testing in high burden settings, and we recently reported improved outcomes in programs in Southern Africa which utilize routine viral load testing.12 South Africa is one of the few high-burden settings to follow the public health approach to ART service delivery, with access to routine viral load monitoring and a standardized approach to confirming virologic failure. In settings with access to viral load testing, WHO recommends switching therapy when a patient has a persistent viral load above 5000 viral copies per milliliter.1 The performance of this guidance at a national level, and the programme-level impact of different thresholds for determining virologic failure have not previously been described.
Individual cohorts have reported on durability of first-line regimens, associations with confirmed virologic failure, and delays in switching to second line, but there is limited understanding of factors predicting switch after virologic failure and how consistent practice is with respect to switching failing patients to second-line therapy.12 We examined the impact of the definition of virologic failure on failure rates, quantified rates of virologic failure and switch to second-line ART and studied predictors of each in a combined cohort of nearly 20,000 patients initiating ART throughout South Africa. We also explored variability in switching rates by treatment program.
Study Design and Population
The data for this analysis come from the international epidemiological databases to evaluate AIDS Southern Africa (IeDEA-SA) collaboration (http://www.iedea-sa.org/), a collaboration of HIV treatment programs in Southern Africa combining prospectively collected clinical treatment data from 25 programs.23 At the time of data transfer, participating sites represented nearly 10% of all South African adult public-sector patients initiating ART.24 We included data from 5 programs, operating throughout the country, with adequate viral load data. All clinics followed the 2004 South African National Treatment Guidelines for public sector ART provision.25 Although these guidelines have recently been revised,26 during the study period patients whose CD4 cell count declined to below 200 cells per cubic milliliter or who were diagnosed with a WHO stage 4 condition (excluding extra-pulmonary tuberculosis) were eligible for ART. Standard first-line regimens comprised stavudine (d4T) with lamivudine (3TC), and either efavirenz (EFV), or nevirapine (NVP). Zidovudine (AZT) was available for specific indications. Once on ART, guidelines provided for 6-monthly viral load and CD4 count monitoring. In 4 of the 5 sites, viral load testing was done by the National Health Laboratory Service (NucliSens EasyQ HIV-1 assay; bioMérieux, Boxtel, The Netherlands), whereas in Gugulethu, a private laboratory provided the viral load testing (Bayer HIV-1 RNA 3.0 assay, Leverkusen, Germany). Virological failure was defined as a 2 consecutive viral loads above 400 copies per milliliter, with the second value >5000 copies per milliliter despite stepped-up adherence interventions. The recommended second-line regimen was AZT, didanosine (ddi), and lopinavir–ritonavir (LPVr). There was some variation in the application of these guidelines, largely because some sites initiated ART prior to formalization of national guidelines. This included using AZT over d4T, the timing of the first monitoring viral load, the frequency of monitoring and the sequential thresholds used to designate confirmed virologic failure (Table 1). Sites nevertheless closely followed a uniform national programme with respect to ART eligibility, patient preparation, regimens, monitoring frequency and failure definitions,27 facilitating this combined analysis.
Eligible patients were ART naïve adults aged ≥16 initiated on a standard first-line ART regimen at one of the 5 sites between January 2000 and July 2008, and who completed ≥6 months of follow-up after ART initiation. We conducted 3 analyses: of 1) time to virologic treatment failure in which we varied the failure definition; 2) the rate and predictors of treatment failure using a common virologic failure definition; and 3) time to and predictors of switching to second-line ART among those failing first-line. For the time-to-switch analysis we included only patients in care for six-months after documented virologic failure. Patients who were never virologically suppressed were included in the main analyses, but were excluded in a sensitivity analysis. Viral load testing completeness at 2 years duration on ART was estimated for each cohort for all patients followed up for at least 27 months based on viral loads taken between 21 and 27 months on ART.
For analysis 1, we defined treatment failure as a detectable viral load (≥400 copies/mL) after 6 months on first-line ART followed by a second consecutive viral load above a threshold (which varied) separated by between 2 weeks and 1 year without suppression during that time. We varied the threshold for the second elevated value (≥400, ≥1000, ≥5000, ≥10,000 copies/mL). Patients who died after a first elevated viral load (N = 79) with no further elevated viral load were not considered failures. Because some patients who were likely virologic failures died or left care before receiving a confirmatory viral load, in sensitivity analyses we analyzed rates of confirmed failure or death after initial viral load elevation to account for missed failures in patients who died. For analysis 2 we used a threshold of ≥1000 (referred to as the “common failure definition”), because no cohort had a protocol requiring switching below this threshold, which is now used in the South African national ART programme.28 Analysis 3 was limited to patients meeting their clinic's failure definition (Table 1), because sites typically would not switch patients who had not met their definition. We defined switching to second-line ART as initiating a protease inhibitor with a change in at least one nucleoside reverse transcriptase inhibitor, ≥6 months after starting first-line therapy.
We defined treatment gaps as the number of days between the end date of all drugs in one prescribed regimen and the start date of the next regimen. We defined CD4 count at ART initiation as the last measure 12 months before through 14 days after initiation (87% of patients had such a measure). We defined viral load at ART initiation as the last viral load measure 6 months before through 3 days after ART initiation. For predictors of switching, CD4 count and viral load at failure were defined as the temporally last measures 3 months before through 2 weeks after the second detectable failing viral load. Because we hypothesized clinicians would prioritize switching failing patients with low or dropping CD4 counts, we include time updated “current” CD4 count after failure and CD4 decline as predictors in multivariable models. For missing current CD4 counts, we carried forward the last measure up to 9 months.
For failure analyses, person-time accrued from six months after initiating ART until the earliest of death, loss to follow-up (LTFU), 5 years on ART, administrative censoring (which varied by cohort but at latest was Jan 2009), or treatment failure (date of the second elevated viral load). For switching analyses, person-time accrued from treatment failure until the earliest of death, LTFU, administrative censoring, or switch. LTFU was determined to have occurred at the last recorded visit and was defined based on each of the individual sites' definitions.
We derived Kaplan-Meier estimates of cumulative failure probabilities using the differing failure definitions and stratified by predictors of failure. We estimated hazard ratios (HR) for associations of patient characteristics with both virologic treatment failure and switch to second-line ART using Cox proportional hazards regression. In addition to age and sex, we included variables with a univariate P-value <0.2 in multivariable models. For failure models we included cohort, ART initiation year, nonnucleoside reverse transcriptase inhibitor (NNRTI) in first-line ART (EFV/NVP), TB treatment (Yes/No/Missing), WHO stage (I/II, III/IV, missing), CD4 count and viral load at ART initiation and treatment interruptions of 7 days or more. For switch models we included cohort, year of failure, years on ART at failure, current CD4 count, % change in CD4 count from failure, and viral load at failure. As a sensitivity analysis, missing baseline values were also imputed using fully conditional modeling by means of a chained equations approach,29 with estimation results combined by Rubin's rules.30
Approval for analyses was given by the Universities of Cape Town, Bern and Boston University. All sites had ethical approval to transfer anonymized data to the IeDEA data center.
Details of the 5 public-sector treatment sites in South Africa are given in Table 1. While some patients initiated treatment as early as 1999, the majority initiated treatment since the public-sector ART rollout in South Africa began in April 2004. Viral load testing completeness at 2 years on ART ranged from 71% to 85%.
Of 23,465 adults completing 6 months of ART, 3,820 were excluded because they initiated a non-standard first-line regimen or their first-line regimen could not be determined. The 19,645 eligible patients were followed for 29,935 person-years (py) (median (range across sites) 1.3 years (1.1–1.4) in the study and 1.8 years on ART). Excluded patients had similar age, gender, cohort, TB treatment at ART initiation to eligible patients, but were more likely to have initiated ART before 2004 (19% vs. 5%), be WHO Stage I/II (47% vs. 38%) and have CD4 count >350 at ART initiation (27% vs. 1%).
Two programs accounted for 73% of all patients (Table 2). The majority of patients were female (66%), initiated ART after 2005 (>51%), and initiated d4T-3TC-EFV (68%) or d4T-3TC-NVP (21%). Advanced immunosuppression was common at ART initiation: 62% of patients were WHO Stage III/IV, 30% had a CD4 count <50 and 36% were on tuberculosis treatment. Median CD4 count at ART initiation was 93 (IQR:39–155) cells/μL. Of 19,456 patients with known outcome 79.0% were alive and in care, 4.7% died, 9.4% were lost to follow-up and 6.9% transferred.
Of the 19,645 patients, 17,272 (88%) achieved virologic suppression on first-line ART and 1348 (9.9%, 4.5/100 person-years) met the common failure definition (threshold ≥1000). The median time from ART initiation until treatment failure among those who failed was 16 months (IQR: 12–23), while the median time between the first and second detectable viral loads was 2.7 months (IQR: 1.6–4.7).
First-line Failure by Viral Load Threshold
Figure 1 shows cumulative probabilities of virologic treatment failure, using different thresholds for confirmation in a Kaplan-Meier survival analysis. Between 6 months and 5 years on ART, using our most sensitive confirmatory threshold (≥400 copies/mL), 16.9% [95% confidence interval (CI): 15.4% to 18.6%] of patients failed. The proportion failing was similar when increasing this threshold to 1000 copies per milliliter, but was substantially less using thresholds of 5000 (10.0%; 95% CI: 8.8% to 11.2%) or 10,000 copies per milliliter (7.8%; 95% CI: 6.6% to 9.3%, or 2.2 times less likely to meet the definition compared to a threshold of 400 copies/mL). This last estimate should be interpreted with caution because no cohort used a threshold of 10,000 copies per milliliter, and so some patients failing at lower thresholds will have switched before reaching a threshold of 10,000. When we limited the analysis for each threshold to only those cohorts which in practice defined failure as greater than or equal to the threshold examined and exclude the 10,000 group, results changed little (failure was 1.5 times greater comparing a threshold of 400 to a threshold of 5000). In sensitivity analyses, including deaths after a single elevated viral load as failures, cumulative failure probability at 5 years on ART ranged from 16.7% to 24.7% depending on the definition used.
Predictors of First-Line Treatment Failure
Table 3 shows associations of patient characteristics with treatment failure, using the common definition. In multivariable models (Table 3), age [per 10 year increase adjusted hazard ratio (aHR): 0.73; 95% CI: 0.68 to 0.79], NVP use (aHR: 1.45; 95% CI: 1.26 to 1.66), and treatment provider were associated with failure. The association between NVP use and failure remained when modeling this separately for men and women (aHR: 1.70, 95% CI: 1.30 to 2.21; 1.33, 1.13 to 1.56). One cohort had a substantially lower failure rate compared with the others. A CD4 count <25 at ART initiation cells per microliter was associated with a 60% increased risk of failure compared with those 100–199 cells per microliter (aHR: 1.60; 95% CI: 1.37 to 1.87). Although gaps in treatment of 7 days or more were uncommon (n = 912, 4.6% of patients), 2 or more gaps was associated with a 7-fold increased risk of failure (aHR: 7.25; 95% CI: 4.95 to 10.6), and one gap associated with a 2-fold increased risk (aHR: 2.46; 95% CI: 2.11 to 2.87). Results were very similar when using multiple imputation for missing data and when limited to patients who initially achieved viral suppression. Results were also similar for the outcome of failure or death except for an attenuation in the association with NVP in the initial regimen, the increased risk associated with 2 or more gaps in care was reduced (HR: 2.74; 95% CI: 2.19 to 3.45) and a stronger association between later year of ART initiation and reduced failure/death was observed (data not shown).
Switching to Second Line
Among the 1348 patients meeting the common failure definition, 62% (833/1348) switched to second-line ART. Overall, 10.1% of the cohort were switched (9.0%–11.4%) to second line between 6 months and 5 years on ART (Fig. 1). Of those who completed at least 6 additional months of follow-up, 664 (74%) switched at some point after failure. Those who switched did so a median of 4.6 months after failure (IQR: 2.1–8.7). The majority (81%) were switched to the government-recommended regimen of AZT-ddi-LPVr.
Table 4 displays associations of patient characteristics with switching among those who failed according to their site's failure definition. In a multivariable analysis, CD4 count at failure was weakly associated with switching. Those with a higher rate of decline in CD4 count since failure were however more likely to switch. A CD4 % drop of –100% to –51% versus remaining relatively stable (–25% to 25%) was associated with a 2-fold increased switch rate (aHR: 1.82; 95% CI: 1.24 to 2.68) although a CD4 % drop of −50% to 0% compared with remaining stable (–25% to 25%) was associated with a 1.3-fold increased switch rate (95% CI: 0.93 to 1.77). There remained a 2-fold difference between the cohorts with the fastest and slowest times to switching in virologically failing patients.
In one of the largest studies to date from the South African national treatment program exploring the extent of virologic treatment failure, we found 8%–17% of patients failed first-line therapy by 5 years on treatment using survival analysis depending on the definition of confirmed failure used in line with findings from individual cohorts.2,6 There were expected delays both between a first elevated viral load and confirmation of treatment failure and subsequent switching of therapy in those who switched (median of 2.7 and 4.6 months respectively). Although nearly 3 quarters of patients with confirmed virologic failure and at least 6 months of additional follow-up had switched, there was up to a 2-fold difference in time to switching between cohorts, with switching occurring faster in patients with rapidly falling CD4 counts.
Definition of Virologic Failure
South African and WHO guidelines suggest a pragmatic approach to defining virologic treatment failure, which relies on confirmation of viremia after attempts to optimize adherence. The World Health Organization 2010 guideline revisions recommend that when viral load monitoring is available, “A persistent viral load of >5000 copies per milliliter confirms treatment failure.”1 South African guidelines in operation at the time of this study used the same confirmatory threshold, but allowed for the first elevation to be above 400 copies per milliliter.
The approach has been shown to successfully select out-patients with high levels of resistance warranting switching to second-line therapy, but low levels of cross-resistance between first and second-line regimens.31 As evidenced by this study however, the exact interpretation of this approach is varied and can profoundly impact the number of patients who meet the failure definition and require second-line therapy.
Three subsequent guideline changes in South Africa may impact the interpretation of these findings.28 In April 2010, the initial recommended nucleoside reverse transcriptase inhibitor backbone was changed from d4T and 3TC to tenofovir and 3TC, whereas routine viral load testing frequency was dropped from 6 monthly to annually beyond the first year on ART. The threshold for confirming virologic failure was lowered from 5000 to 1000 copies per milliliter with the confirmatory test now required within 3 months of the initial elevation.
Whereas the less frequent monitoring may increase delays to identifying patients failing virologically, with the introduction of tenofovir there may also now be fewer concerns about the accumulation of thymidine analogue mutations while viraemic, with the associated potential to compromise second-line therapy.
Switching of Failing Patients
A previous analysis from the IeDEA collaboration demonstrated switching occurred more frequently and at higher CD4 counts in sites with viral load monitoring.32 Mortality was lower and CD4 trajectories steeper in viral load sites,12 though this has not been consistent across studies.33 The current study provides further insight into the period between virologic failure confirmation and switching. A high proportion of patients who should be switched are switched and compliance with guideline advice is more complete than is the case for children in South Africa.34 Nevertheless, the median delay of nearly 5 months between confirmation of failure and switch combined with intercohort differences suggests that administrative and clinical factors are additionally impacting on compliance with switching guidelines. The strong association between CD4 count trajectory and switching further suggests that clinical judgment is a contributor to this variability.
Current data suggest protease inhibitor–based second-line therapy is not being compromised by delays in switching resulting from South Africa's pragmatic virologic failure guidelines, with the majority of patients failing second-line therapy remaining susceptible to boosted lopinavir.35,36 The finding that most early failures on second-line are adherence related supports the provision for a period of active adherence optimization before switching. Delays in switching on the other hand place patients at increased risk of illness and death through longer durations spent viremic and at lower CD4 counts.12,13 An important follow-on analysis will therefore be to estimate the causal effect of delays in switching patients with confirmed virologic failing.37
Associations With Failure
The associations with virologic failure we found mirror those found in individual cohorts. Measures of advanced disease (CD4 count, WHO stage, viral load) were associated with failure. We found a strong association with treatment interruptions, perhaps serving both as a proxy for poor adherence and as a consequence of the long half-life of NNRTIs which remain in circulation longer than the other drugs following unplanned interruptions.38,39 We again found NVP as choice of NNRTI was associated with virologic failure, a consistent finding across observational studies from different settings2,40–42 and is not in conflict with the clinical trial data for African sites.43,44
Our study has several limitations. First, we lacked good PMTCT data to be able to examine the role of single-dose NVP exposure in treatment failure. Second, confounding by indication could have occurred, particularly in the relationship between NVP use and failure in sites where EFV was more commonly used. Next, although we observed modest differences between cohorts in the completion and frequency of viral load testing, failure to test or report viral load results according to guidelines would reduce the probability of meeting failure definitions. We also found the overall failure rate was sensitive to assumptions made about whether patients with only a single detectable viral load before death were truly failures. Also differences in how failure was defined by each cohort could have led to underestimates of failure rates using definitions with higher thresholds if patients with detectable viral loads below the threshold were switched before they could reach a higher threshold. Differences in follow-up time between cohorts may also explain some of the differences in failure rates observed. Finally, we had no data on 2 important potential predictors of treatment failure, adherence45 and prior resistance.
In conclusion, future treatment guidelines revisions should make explicit the rationale for the thresholds chosen to define and confirm virologic failure in light of our finding that these profoundly on the proportion of patients who meet failure definitions, and resultant costs of second-line treatment. Although guidance on switching failing patients is generally followed, there remains considerable variability in time to switching after failure, due to both clinician and administrative factors. Future studies should investigate the impact failure definitions and delays in switching have on subsequent treatment outcomes.
The authors thank all the patients whose data were used in this analysis. The authors also thank all staff at participating sites for preparation of data contributed to the IeDEA Southern Africa collaboration. Many thanks to Nicola Maxwell for preparing the combined data for analysis, to Morna Cornell and Claire Graber for project management and to Michael Schomaker for advice and technical assistance with the analysis. Matthew Fox had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
1. World Health Organization. Antiretroviral Therapy for HIV Infection in Adults and Adolescents Recommendations for a Public Health Approach, 2010 Revision. Geneva, Switzerland: World Health Organization; 2010.
2. Boulle A, Van Cutsem G, Hilderbrand K, et al.. Seven-year experience of a primary care antiretroviral treatment programme in Khayelitsha, South Africa. AIDS. 2010;24:563–572.
3. Stringer JS, Zulu I, Levy J, et al.. Rapid scale-up of antiretroviral therapy at primary care sites in Zambia: feasibility and early outcomes. JAMA. 2006;296:782–793.
4. Sanne IM, Westreich D, Macphail AP, et al.. Long term outcomes of antiretroviral therapy in a large HIV/AIDS care clinic in urban South Africa: a prospective cohort study. J Int AIDS Soc. 2009;12:38.
5. Ferradini L, Jeannin A, Pinoges L, et al.. Scaling up of highly active antiretroviral therapy in a rural district of Malawi: an effectiveness assessment. Lancet. 2006;367:1335–1342.
6. Nglazi MD, Lawn SD, Kaplan R, et al.. Changes in programmatic outcomes during 7 years of scale-up at a community-based antiretroviral treatment service in South Africa. J Acquir Immune Defic Syndr. 2011;56:E1–E8.
7. May M, Boulle A, Phiri S, et al.. Prognosis of patients with HIV-1 infection starting antiretroviral therapy in sub-Saharan Africa: a collaborative analysis of scale-up programmes. Lancet. 2010;376:449–457.
8. Fox MP, Ive P, Long L, et al.. High rates of survival, immune reconstitution, and virologic suppression on second-line antiretroviral therapy in South Africa. J Acquir Immune Defic Syndr. 2010;53:500–506.
9. Pujades-Rodriguez M, O'Brien D, Humblet P, et al.. Second-line antiretroviral therapy in resource-limited settings: the experience of Medecins Sans Frontieres. AIDS. 2008;22:1305–1312.
10. Palombi L, Marazzi MC, Guidotti G, et al.. Incidence and predictors of death, retention, and switch to second-line regimens in antiretroviral- treated patients in sub-Saharan African Sites with comprehensive monitoring availability. Clin Infect Dis. 2009;48:115–122.
11. Renaud-Thery F, Nguimfack BD, Vitoria M, et al.. Use of antiretroviral therapy in resource-limited countries in 2006: distribution and uptake of first- and second-line regimens. AIDS. 2007;21(suppl 4):S89–S95.
12. Keiser O, Chi B, Gsponer T, et al.. Outcomes of antiretroviral treatment in programmes with and without routine viral load monitoring in Southern Africa. AIDS. 2011;25:1761–1769.
13. Keiser O, Tweya H, Braitstein P, et al.. Mortality after failure of antiretroviral therapy in sub-Saharan Africa. Trop Med Int Health. 2010;15:251–258.
14. World Health Organization, UNICEF. Towards Universal Access: Scaling up Priority HIV/AIDS Interventions in the Health Sector—Progress Report, September 2009. Geneva, Switzerland: World Health Organization; 2009.
15. Boyd MA. Current and future management of treatment failure in low- and middle-income countries. Curr Opin HIV AIDS. 2010;5:83–89.
16. Harries AD, Zachariah R, van Oosterhout JJ, et al.. Diagnosis and management of antiretroviral-therapy failure in resource-limited settings in sub-Saharan Africa: challenges and perspectives. Lancet Infect Dis. 2010;10:60–65.
17. Keiser O, MacPhail P, Boulle A, et al.. Accuracy of WHO CD4 cell count criteria for virological failure of antiretroviral therapy. Trop Med Int Health. 2009;14:1220–1225.
18. van Oosterhout JJ, Brown L, Weigel R, et al.. Diagnosis of antiretroviral therapy failure in Malawi: poor performance of clinical and immunological WHO criteria. Trop Med Int Health. 2009;14:856–861.
19. Moore DM, Awor A, Downing R, et al.. CD4+ T-cell count monitoring does not accurately identify HIV-infected adults with virologic failure receiving antiretroviral therapy. J Acquir Immune Defic Syndr. 2008;49:477–484.
20. Mee P, Fielding KL, Charalambous S, et al.. Evaluation of the WHO criteria for antiretroviral treatment failure among adults in South Africa. AIDS. 2008;22:1971–1977.
21. Bisson GP, Gross R, Strom JB, et al.. Diagnostic accuracy of CD4 cell count increase for virologic response after initiating highly active antiretroviral therapy. AIDS. 2006;20:1613–1619.
22. Reynolds SJ, Nakigozi G, Newell K, et al.. Failure of immunologic criteria to appropriately identify antiretroviral treatment failure in Uganda. AIDS. 2009;23:697–700.
23. Egger M, Ekouevi DK, Williams C, et al.. Cohort profile: the international epidemiological databases to evaluate AIDS (IeDEA) in sub-Saharan Africa. Int J Epidemiol. [published online ahead of print May 18, 2011] doi:10.1093/ije/dyr080.
24. Cornell M, Grimsrud A, Fairall L, et al.. Temporal changes in programme outcomes among adult patients initiating antiretroviral therapy across South Africa, 2002–2007. AIDS. 2010;24:2263–2270.
25. National Department of Health, Republic of South Africa. South African National Antiretroviral Treatment Guidelines. 1st ed. Pretoria, South Africa: National Department of Health; 2004.
26. National Department of Health, Republic of South Africa. Clinical Guidelines for the Management of HIV&AIDS in Adults and Adolescents. Pretoria, South Africa: National Department of Health; 2010.
27. Cornell M, Technau K, Fairall L, et al.. Monitoring the South African National Antiretroviral Treatment Programme, 2003-2007: the IeDEA Southern Africa collaboration. S Afr Med J. 2009;99:653–660.
28. National Department of Health Republic of South Africa. Clinical Guidelines for the Management of HIV&AIDS in Adults and Adolescents. Geneva, Switzerland: World Health Organization; 2010.
29. van Buuren S, Boshuizen HC, Knook DL. Multiple imputation of missing blood pressure covariates in survival analysis. Stat Med. 1999;18:681–694.
30. Rubin DB. Multiple imputation after 18+ years. J Am Stat Assoc. 1996;91:473–489.
31. Orrell C, Walensky RP, Losina E, et al.. HIV type-1 clade C resistance genotypes in treatment-naive patients and after first virological failure in a large community antiretroviral therapy programme. Antivir Ther. 2009;14:523–531.
32. Keiser O, Tweya H, Boulle A, et al.. Switching to second-line antiretroviral therapy in resource-limited settings: comparison of programmes with and without viral load monitoring. AIDS. 2009;23:1867–1874.
33. Mermin J, Ekwaru JP, Were W, et al.. Utility of routine viral load, CD4 cell count, and clinical monitoring among adults with HIV receiving antiretroviral therapy in Uganda: randomised trial. BMJ. 2011;343:D6792.
34. Davies MA, Moultrie H, Eley B, et al.. Virologic failure and second-line antiretroviral therapy in children in South Africa—the IeDEA Southern Africa collaboration. J Acquir Immune Defic Syndr. 2011;56:270–278.
35. Hosseinipour MC, Kumwenda JJ, Weigel R, et al.. Second-line treatment in the Malawi antiretroviral programme: high early mortality, but good outcomes in survivors, despite extensive drug resistance at baseline. HIV Med. 2010;11:510–518.
36. Wallis CL, Mellors JW, Venter WD, et al.. Protease inhibitor resistance is uncommon in HIV-1 subtype C infected patients on failing second-line lopinavir/r-containing antiretroviral therapy in South Africa. AIDS Res Treat. 2011;2011:769627.
37. Hernan MA, Lanoy E, Costagliola D, et al.. Comparison of dynamic treatment regimes via inverse probability weighting. Basic Clin Pharmacol Toxicol. 2006;98:237–242.
38. Fox Z, Phillips A, Cohen C, et al.. Viral resuppression and detection of drug resistance following interruption of a suppressive non-nucleoside reverse transcriptase inhibitor-based regimen. AIDS. 2008;22:2279–2289.
39. Bansi LK, Benzie AA, Phillips AN, et al.. Are previous treatment interruptions associated with higher viral rebound rates in patients with viral suppression? AIDS. 2008;22:349–356.
40. Nachega JB, Hislop M, Dowdy DW, et al.. Adherence to nonnucleoside reverse transcriptase inhibitor-based HIV therapy and virologic outcomes. Ann Intern Med. 2007;146:564–573.
41. Keiser P, Nassar N, White C, et al.. Comparison of nevirapine- and efavirenz-containing antiretroviral regimens in antiretroviral-naive patients: a cohort study. HIV Clin Trials. 2002;3:296–303.
42. Cozzi-Lepri A, Phillips AN, d'Arminio Monforte A, et al.. Virologic and immunologic response to regimens containing nevirapine or efavirenz in combination with 2 nucleoside analogues in the Italian Cohort Naive Antiretrovirals (I.Co.N.A.) study. J Infect Dis. 2002;185:1062–1069.
43. van Leth F, Phanuphak P, Ruxrungtham K, et al.. Comparison of first-line antiretroviral therapy with regimens including nevirapine, efavirenz, or both drugs, plus stavudine and lamivudine: a randomised open-label trial, the 2NN Study. Lancet. 2004;363:1253–1263.
44. Wester CW, Thomas AM, Bussmann H, et al.. Non-nucleoside reverse transcriptase inhibitor outcomes among combination antiretroviral therapy-treated adults in Botswana. AIDS. 2010;24(suppl 1):S27–S36.
45. Ramadhani HO, Thielman NM, Landman KZ, et al.. Predictors of incomplete adherence, virologic failure, and antiviral drug resistance among HIV-infected adults receiving antiretroviral therapy in Tanzania. Clin Infect Dis. 2007;45:1492–1498.
STUDY PROFILE OF VIROLOGIC TREATMENT FAILURE AND SWITCHING TO SECOND-LINE ANTIRETROVIRAL THERAPY IN THE IEDEA-SA SOUTH AFRICA COHORT
KAPLAN–MEIER CURVES OF PREDICTORS OF FAILURE STRATIFIED BY A) COHORT, B) CD4 COUNT AT ART INITIATION, C) VIRAL LOAD AT ART INITIATION, AND D) GAPS IN TREATMENT
Log-rank P value for cohort (P < 0.0001), CD4 count at ART initiation (P < 0.0001), viral load at ART initiation (P = 0.0011), and gaps in treatment (P < 0.0001).
IEDEA SOUTHERN AFRICA STEERING GROUP AND SITE PRINCIPAL INVESTIGATORS
Cape Town, South Africa: Brian Eley, Red Cross Children's Hospital; Daniele Garone, Khayelitsha ART Programme and Médecins Sans Frontières; Robin Wood, Gugulethu and Masiphumelele ART Programmes and Desmond Tutu HIV Centre; Hans Prozesky, Tygerberg Academic Hospital. Johannesburg, South Africa: Christopher Hoffmann, Aurum Institute for Health Research; Patrick MacPhail, Themba Lethu Clinic, Helen Joseph Hospital; Harry Moultrie, Wits Reproductive Health and HIV Institute; Karl Technau, Rahima Moosa Mother and Child Hospital. KwaZulu-Natal, South Africa: James Ndirangu, Hlabisa HIV Treatment and Care Programme; Janet Giddy, McCord Hospital, Durban. Zimbabwe: Cleophas Chimbetete, Newlands Clinic, Harare; Christiane Fritz, SolidarMed Zimbabwe. Malawi: Sam Phiri, Lighthouse Clinic, Lilongwe. Mozambique: Sabrina Pestilli, SolidarMed Mozambique; Paula Vaz, Paediatric Day Hospital, Maputo. Zambia: Jeff Stringer, Center for Infectious Disease Research. Cited Here...