When does viral load testing add value in sub-Saharan Africa?

Routine viral load testing did not catch failure of first-line treatment earlier than discretionary viral load testing, but it did reduce the period on failing treatment by nearly a year in a study conducted in Zambia, Dr Mike Saag of the University of Alabama at Birmingham reported to the 19th Conference on Retroviruses and Opportunistic Infections in Seattle this week.

Viral load testing remains a luxury for many antiretroviral treatment (ART) programmes in low- and middle-income countries owing to its high cost and the sophisticated laboratory equipment that is needed to carry out a viral load test. However, viral load testing is being adopted and expanded in some countries as laboratory capacity is upgraded, primarily as a confirmatory test in cases where patients appear to be experiencing treatment failure.

Mike Saag highlighted the dilemma faced when Zambia began planning its antiretroviral roll-out in 2004: should the programme spend nearly $100 per patient per year on viral load testing, or seek to treat more patients at a cost of $240 without viral load monitoring? Inevitably the need to treat as many people as possible won out, but as treatment programmes mature and viral load testing becomes more accessible, many clinicians and programme managers are asking: what is the best strategy for using viral load testing? Can it prevent unnecessary switches to second-line treatment, or reduce the risk of high-level resistance that destroys future treatment options?

To evaluate whether routine viral load testing would add value in Zambia’s PEPFAR-supported programme, the Center for Infectious Disease Research of Zambia designed a cluster randomised trial which compared discretionary ('standard of care') viral load testing of patients on antiretroviral treatment with testing at regular intervals.

Clinics were selected and paired with control clinics on the basis of similar mortality rates in the previous year, in order to minimise any bias due to variations in death rates between clinics.

Discretionary testing was carried out in the standard-of-care clinics if patients met any of the following clinical criteria suggestive of treatment failure:

• New or recurrent WHO stage III / IV or failure to improve clinically after 6 months on ART
• Less than 50 cells/mm³ increase in CD4 count at 6 months of therapy in the absence of marked clinical improvement; CD4 < 100 cells/mm³ after 12 months on treatment
• More than 30% decline in CD4 count from its peak post-treatment initiation value
• Decline in CD4 count to a level below that of treatment initiation after 6 months of treatment.

In the routine testing clinics, viral load testing was carried out at fixed time points: three months after starting treatment, at six months and then every six months.

In both groups a confirmatory test was carried out if a viral load result was returned above 400 copies/ml to determine whether the case was a viral load 'blip' that was subsequently resuppressed, or a genuine case of virologic rebound.

The primary outcome of the study was the relationship between viral load testing modality and mortality, but as Dr Saag pointed out, the statistical power of the study to measure differences in mortality between sites was rapidly diminished after one year of treatment availability due to the dramatic reduction in death rates at all sites.

The secondary endpoints of the study are the proportions of patients who switch to second-line therapy, and the proportion who develop resistance after first-line treatment failure according to viral load testing protocol.

The study enrolled 1973 patients starting antiretroviral therapy between December 2006 and May 2008.

Participants initiating therapy at the participating clinics were well matched, with similar viral load levels (around 100,000 copies/ml), CD4 cell counts (around 145 cells/mm³) and similar proportions with WHO stage III or IV disease (67% and 69%).

There was no significant difference in overall mortality according to the clinic’s viral load testing protocol after one year, and a similar proportion of patients who experienced rebound above 400 copies/ml had returned to a viral load below 400 copies/ml at their next clinic visit (48%).

Sustained virologic response (<400 copies/ml) was seen in 82% of the routine viral load group and 90% of the standard-of-care group, a non-significant difference. Furthermore there was no significant difference in the time to virologic failure (371 days in the routine testing group and 399 days in the standard-of-care group).

However a significant difference did emerge when the rate of switching to second-line treatment, and the delay between viral rebound and switching to second-line treatment, were compared.

Patients in the routine testing clinics had almost twice the rate of switching to second-line treatment (hazard ratio 1.94, 95% confidence interval 1.22-3.10). The median time to a switch to second-line treatment was 168 days in patients with virologic failure at routine testing clinics, but 560 days at the discretionary testing clinics.

The researchers note that shortage of second- and third-line treatment options may have delayed switching, and more analyses on the standard-of-care group need to be carried out.

Cost-effectiveness and resistance analyses have still to be completed, but the authors concluded that, at the very least, routine viral load testing clearly limits the amount of time spent on a failing regimen, and the results of the resistance analysis will provide important information on the costs of that delay for future treatment options.