University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

Tenofovir in second-line ART in Zambia and South Africa: Collaborative analysis of cohort studies
Global Health Sciences Literature Digest
Published September 30, 2012
Journal Article

Wandeler G, Keiser O, Mulenga L, Hoffmann CJ, Wood R, Chaweza T, Brennan A, Prozesky H, Garone D, Giddy J, Chimbetete C, Boulle A, Egger M; for the IeDEA Southern Africa Collaboration. Tenofovir in second-line ART in Zambia and South Africa: Collaborative analysis of cohort studies. J Acquir Immune Defic Syndr. 2012 Sep 1; 61(1):41-48.


To compare outcomes of patients receiving tenofovir (TDF)-containing second-line antiretroviral therapy (ART) with those on second-line regimens not containing TDF.


Six clinics and ART programs in South Africa (n=5) and Zambia (n=1).

Study Design

Retrospective analysis of observational cohort data in the International epidemiological Databases to Evaluate AIDS in Southern Africa (IeDEA-SA) network.


Adults (age ≥16) receiving second-line ART.

Main Outcome Measures

Time to immunologic failure, time to virologic failure, and time to death.


The IeDEA-SA network is a regional collaboration of ART programs. In each program, patient data are collected at ART initiation and at each follow-up visit, using standardized instruments, and transferred to data centers at universities in South Africa and Switzerland. Investigators included all cohorts with >50 patients on second-line ART, ≥10 patients on TDF-containing second-line regimens and ≥10 patients on non-TDF-containing second-line regimens.

Characteristics at the start of second-line ART were compared between patients on second-line regimens containing and not containing TDF. Investigators compared rates of immunologic and virologic failure in Cox regression models, measuring time from six months after the patient switched to second-line ART. In South Africa, viral load and CD4 cell counts are monitored every six months during the first year of ART and then annually. In Zambia, CD4 counts are monitored every six months; viral load measurements are not routinely performed. Researchers used competing risk cumulative incidence curves and competing risk regression models to compare mortality, measuring time from when the patient switched to second-line ART.

All regression models included the variables gender, age, CD4 cell count at the start of second-line ART, time on a first-line regimen before switching to second-line ART and calendar year of starting second-line ART. Analyses were performed separately for each of the two countries.

Finally, to assess the effect of the first-line backbone on second-line outcomes, researchers also examined whether the use of stavudine (d4T) in the first-line ART regimen predicted immunologic failure in patients on a TDF-containing second-line regimen.


A total of 3243 patients on second-line ART were included in the analyses, including 1556 (48.0%) on a TDF-containing regimen. In South Africa, 206/1556 (13.2%) patients were on TDF-based second-line ART with proportion ranging from 4% to 25% among the five sites. In Zambia, 1350/1687 (80%) of patients were on TDF-based second-line ART. Nearly all patients (3225/3243, 99.4%) were on second-line regimens containing ritonavir-boosted lopinavir (LPV/r). Most patients (1468/1687, 87.0%) on a non-TDF-containing second-line regimen had a backbone of didanosine (ddI) plus zidovudine (ZDV) or ddI plus abacavir (ABC).

Over 4471 person-years, 127 (3.9%) patients died. Patients on TDF-containing regimens were at a marginally lower risk of dying than those on non-TDF regimens (hazard ratio [HR] 0.63, 95% confidence interval [CI] 0.38 to 1.05). In contrast to Zambia, however, the use of TDF-containing regimens was not associated with reduced mortality in South Africa (HR 1.13, 95% CI 0.45 to 2.79). At 3 years, 3.3% (95% CI 2.3% to 4.5%) of patients in the TDF group in Zambia and 4.4% (95% CI 1.8 to 8.8) in South Africa were known to have died. These proportions were higher in the non-TDF groups: 9.0% (95% CI, 6.9% to 12.7%) in Zambia and 7.8% (95% CI 5.8 to 10.00) in South Africa.

Patients on TDF-containing regimens had significantly lower risks of immunologic failure (HR 0.60, 95% CI 0.41 to 0.87) and virologic failure (HR 0.28, 95% CI 0.09 to 0.90) than those on non-TDF regimens. Over 2782 person-years, 94 (7.9%) patients on TDF-containing regimens, and 146 (12.8%) patients on non-TDF regimens developed immunologic failure. The crude incidence rate of immunologic failure was 69.9 (95% CI 57.1 to 85.6) per 1000 person-years in the TDF group and 101.6 (95% CI 86.4 to 119.4) per 1000 person-years in the non-TDF group. In South Africa, three patients (2.7%) in the TDF group and 107 patients (12.1%) in the non-TDF group experienced virologic failure (HR 0.28, 95% CI 0.09-0.90).

The risk of treatment failure and death on second-line ART containing TDF and not containing in Zambia and South Africa.

In both countries, male patients and those age <30 were more likely to fail treatment. Time spent on first-line ART before switching to a second-line regimen did not affect outcomes. The risk of immunologic failure in patients on TDF-containing second-line regimens was slightly higher if d4T were used in the first-line backbone but did not reach statistical significance (HR 1.30, 95% CI 0.84 to 2.02).


The authors conclude that patients on TDF-containing second-line ART were less likely to develop treatment failure in all cohorts and in Zambia, less likely to die, than patients on other regimens. They suggest several possible reasons why mortality did not decrease in South Africa: perhaps the result of differences in the South African health system's capacity compared with that of Zambia; the result of differences in ascertainment of deaths and tracing of patients lost to follow-up; the result of confounding by indication (the relatively few patients who were prescribed TDF in South Africa before 2010 might have been a selected group of sicker patients); or perhaps the result of chance.

Risk of Bias

The overall risk of bias in this study is moderate. Apart from a lack of randomization and the heterogeneity among the treatment sites, differences between settings in background mortality, monitoring and treatment strategies may have biased the results, along with differences between health systems and confounding by indication. However, although the proportion of patients on a TDF-containing second-line regimen varied widely across countries and calendar time, reflecting national treatment guidelines, the association of TDF with reduced rates of treatment failure was consistent within countries and cohorts.

In Context

TDF is increasingly being used in first-line and second-line ART, as well as in pre-exposure prophylaxis (PrEP). In the current World Health Organization (WHO) guidelines on adult and adolescent ART,(1) a TDF-containing regimen is the preferred second-line regimen in patients whose first-line regimens had included d4T or ZDV. In patients whose first-line regimen contained TDF, the preferred second-line regimen is ZDV + lamivudine (3TC) + (atazanavir [ATV]/r or LPVr). New WHO guidelines on ART are expected to be released in 2013.

Programmatic Implications

Clinicians should follow current WHO guidelines in prescribing second-line ART.


  1. World Health Organization. Antiretroviral Therapy for HIV Infection in Adults and Adolescents: Recommendations for a Public Health Approach. 2010 Revision. [accessed 21 September 2012]