Second-line ART in South Africa shows good results

High rates of increased CD4 cell counts and viral suppression, together with low mortality were seen in adults at a large HIV public-sector urban clinic in Johannesburg, South Africa after one year on second-line antiretroviral therapy, Matthew P. Fox and colleagues report in a study published online ahead of print in the Journal of Acquired Immune Deficiency Syndromes.

In 2007 an estimated 300,000 people were receiving antiretroviral therapy in South Africa as a result of the government’s 2004 large-scale plan for the provision of ART. Outcomes for those on first-line regimens have been good. However, there is a growing concern that the number of people failing on treatment will increase as the numbers starting treatment and time on treatment increase.

Furthermore, use of single-dose nevirapine given to women for the prevention-of-mother-to child transmission leads to the increased probability of developing resistance, and consequent failure of first-line regimens containing nevirapine.

The anticipated increase in demand for second-line regimens in resource-poor settings raises the question of their long-term effectiveness, and in particular, the need to understand who fails second-line treatment and why and why.

The high costs of second-line regimens and concerns about the development of resistance in resource-poor settings where viral load testing is uncommon, underscores the importance of this research, yet scant evidence of the effectiveness of second-line regimens in resource-poor settings exists.

A rise in the need for second-line therapy is anticipated in South Africa in the near future. Treatment guidelines for the public sector support switching to a second-line regimen after two consecutive viral loads over 5000 copies/ml.

A cohort study was undertaken using data from a four-year period (April 2004-June 2008) from all patients 18 years of age or older who had begun a standard public sector second-line regimen of zidovudine, didanosine and lopinavir/ritonovir following a standard triple therapy first-line at the Themba Lethu Clinic (TLC) in Johannesburg, South Africa. The clinic’s protocol for switching is based on toxicity or two consecutive viral loads over 1000 copies/ml.

TLC, one of the largest public HIV clinics in South Africa with nearly 11,000 patients having begun antiretroviral treatment, is located at an urban referral hospital. Since 2006 all care at TLC is given free-of-charge.

Treatment monitoring includes CD4 counts and viral loads at four months after start of a new treatment regimen and then every six months thereafter unless clinical assessments indicate otherwise.

The cohort included 328 patients. Data analysis included survival, immunological and virological outcomes after one year of follow-up. The authors chose three measures of treatment success:

• Alive and in care (AIC), which they defined as not known to have died nor lost to follow-up (missed a scheduled ARV pick up for longer than three months);
• Achieving an undetectable viral load – less than 400 copies/ml and
• Increases in CD4 cell count.

At one year of follow-up 78% (243/313) (95% CI: 73%-82%) were alive and in care. However, reaching and maintaining an undetectable viral load is a more precise measure of long-term treatment effectiveness; 77% (203/262) (95%CI: 72%-82%) had an undetectable viral load at one year.

Time-matched comparisons with those on first-line treatment showed those on second-line were only marginally less likely to be alive and in care after one year (HR 0.84, 95%CI: 0.73-0.97).

The authors note that while their findings are similar to a Médecins sans Frontières (MSF) cohort covering resource-poor countries, a comparison is not really valid. The MSF cohort showed a higher percentage of those alive and in care at one year (86%), but the data included several different second-line regimens, with most participants having been on a first-line regimen of stavudine, lamivudine and nevirapine. Standard first-line treatment in South Africa consists of stavudine, lamivudine and efavirenz.

Even with these considerations the authors stress the positive outcomes for those switched to second-line at one year of follow-up.

Mean CD4 cell count increase at one year was 133 cells/mm³. The authors note this compares favourably with the MSF study and represents a significant immune recovery.

Identification of who fails and why is critical to the ongoing success of second-line regimens. The authors found that patients who were switched before they had a second detectable viral load over 1,000 copies/ml (HR 1.68; 95% CI: 1.08-2.61) were more likely to reach undetectable viral load at the end of one year, as were those who switched for reasons other than non-compliance with first-line treatment (HR 1.83; 95% CI: 1.14-2.93).

The authors highlight that in most resource-poor settings antiretroviral treatment failure is detected by clinical assessment alone. Considerable drug resistance will have already developed. In contrast this cohort underwent regular viral load monitoring, and so switching probably occurred earlier than with clinical assessment alone. The authors suggest that as a result, the degree of resistance acquired is limited.

Patients have fewer thymidine analogue mutations (TAMS) or other nucleoside reverse transcriptase inhibitor mutations and so outcomes on second-line regimens are better. This is relevant within the South African context because standard second-line therapy contains zidovudine (AZT) and so depends on the minimal accumulation of TAMS to be effective, note the authors.

Limitations noted include incorrectly categorising the outcome measures. For example, `alive and in care ` has been used as a substitute for vital status, reflecting the realities of working in a large urban treatment facility. No outcomes can be measured for patients who are lost. The authors point to two studies at clinics in Johannesburg where 27% and 48% of those who had dropped out of care and were able to be traced had died.

The authors note that patients who survive failure of first-line treatment long enough to be switched to second-line may have less of a risk for poor outcomes and death than those eligible for second-line who choose not to start it, or those who died or were lost to follow up before getting it. This may have biased mortality rates of those on second-line over first-line. However, patients who switched treatment were matched to patients on first-line ART for the same amount of time, which should have reduced bias.

The authors highlight that the favourable results may reflect a stricter monitoring policy at TLC compared to the national policy.

The authors suggest further study to determine to what extent these findings can be replicated in different settings, as well as longer follow-up to see if these outcomes can be sustained over the long-term. Nonetheless, on the strength of their findings they advocate that the “provision of second-line treatment to patients who fail their first-line ART should be considered a high priority in resource-poor settings.”