Point-of-care viral load testing potentially cost-effective in southern Africa

Routine point-of-care viral load (POC-VL) tests with detection limits of 1000 copies/ml or below will minimise unnecessary treatment changes and improve the cost-effectiveness of antiretroviral therapy (ART) in sub-Saharan Africa, provided it reduces onward transmission and is supported by intensive adherence counselling for people identified with early virologic rebound, researchers report in the advance online edition of AIDS.

In this mathematical modelling study, using longitudinal data from the Gugulethu and Khayelitsha township ART programmes in Cape Town, South Africa, the researchers for leDEA Southern Africa found POC-VL tests with levels of detection under 1000 copies/ml increased costs because of unnecessary switching to second-line ART and did not affect survival compared to clinical and CD4 monitoring.

Three simulated scenarios looked at the possible benefits of routine POC-VL testing in Zambia, Mozambique and Malawi, where only clinical and CD4 monitoring are available at present.

The cost-effectiveness of POC-VL testing, compared to clinical and CD4 monitoring, was poor when detection of treatment failure alone (Scenario A) was considered. Cost-effectiveness improved slightly when the effects of POC-VL monitoring and early switching on HIV transmission was modelled (Scenario B).

However, in Scenario A routine viral load monitoring would only be cost-effective in Zambia when the cost of both viral load monitoring and ART were minimised. This would also prove to be the case under Scenario B in Mozambique.

This pattern changed in Scenario C with the added assumption that the risk of virologic failure is halved with POC-VL because of better adherence counselling following early detection of virologic rebound and subsequent viral re-suppression prior to the emergence of drug resistance. Cost-effectiveness increased with the decreased need for switches to second-line treatment in Scenario C.

The incremental cost-effectiveness ratio (ICER) of POC-VL testing, assuming a unit cost per test between US$5 and US$20 and detection limits between 1000 and 10,000 copies/ml, ranged from US$960 to US$2500 and from cost-saving to US$2460 for each quality-adjusted life-year (QALY) gained for clinical and CD4 monitoring, respectively.

The test’s cost was the key determinant for cost-effectiveness compared to CD4 monitoring, whereas it was the cost of second-line ART that determined cost-effectiveness when compared to clinical monitoring, the authors note.

“In general, lowering the costs of any POC-VL test and of second-line ART are the most promising strategies to maximise the cost-effectiveness of monitoring ART with POC-VL tests”, the authors write.

ICER represents the difference in cost between one intervention and another for each QALY gained. WHO considers an intervention cost-effective if it is no more than three times the GDP per capita.The cost-effectiveness limits for Zambia, Mozambique and Malawi are US$4242 per QALY, US$1749 per QALY and US$1053 per QALY, respectively.

While routine viral load monitoring to detect virologic failure and inform decisions on switching to second-line ART is standard in resource-rich settings, its use in resource-poor settings is limited. Without access to viral load monitoring, many may stay on a failing regimen, so increasing the risks of HIV transmission and development of multi-drug resistance.

The International epidemiologic Databases to Evaluate AIDS in Southern Africa (leDEA-SA) is a collaboration of ART programmes in seven countries in Southern Africa. The researchers restricted their analyses to data from the Gugulethu and Khayelitsa townships where viral load and CD4 cell counts are measured regularly.

They adapted the leDEA-SA mathematical model of ART, which simulates cohorts of patient who are followed from starting ART until death, to compare QALYs, costs and cost-effectiveness between different monitoring strategies in patients receiving ART.

WHO criteria for clinical and immunological failure were used; five alternative thresholds defined virologic failure: 125; 400; 1000; 5000; or 10000 copies/ml.

Costs included appointments, CD4 and VL measurements and an average of the two most common first-line ART regimens. A range of costs for POC-VL was assumed, based on discussions with experts.

In the three scenarios monitoring strategies included: clinical monitoring; six-monthly CD4 count; and VL monitoring tests with a qualitative POC-VL test and levels of detection of 125, 400, 1000, 5000 or 10,000 copies/ml.

Out of 100 true treatment failures, 5 to 6 were observed over a lifetime with viral load monitoring, compared to 37 and 76 with CD4 and clinical monitoring, respectively.

Nine and five out of 100 patients would switch unnecessarily to second-line ART with clinical and CD4 monitoring, respectively, ranging from 15 to 5 or fewer with limits of detection ranging from 125 copies/ml to 1000 copies/ml or more with viral load monitoring.

While viral load monitoring detected treatment failure more accurately, its effect on survival was minimal: the mean QALYs expected at the start of ART ranged from 12.78 with clinical monitoring to 12.93 with POC-VL monitoring.

The number of new infections was higher with a limit of detection of 125 copies/ml when compared to higher limits of detection. This is due “to the large number of unnecessary switches to second-line, which will move second-line failure forward in time and in the absence of further treatment options, increase the number on failing regimens”.

The authors excluded viral load monitoring strategies using a limit of detection of 125 or 400 copies/ml from the cost-effectiveness analyses because they were more expensive, did not improve survival and caused more HIV transmissions.

In most scenarios viral load monitoring was not cost-effective even when the test was reduced to $5 because of the increased need for second-line ART as well as additional tests to confirm failure. Only in Scenario C was POC-VL monitoring clearly a cost-effective strategy.

The assumed benefits, the authors note, of Scenarios B and C will vary according to the setting, sexual behaviour and adherence interventions.

While the authors’ findings support VL tests with detection limits of 1000 copies/ml, they note this is contrary to current practice.

Limitations include a lack of data on the key question of the effect of VL monitoring and subsequent rate of virologic failure; the authors assumed that without VL monitoring, failure would be twice as high.

This assumption was confirmed in a recent study from Kenya showing routine viral load monitoring almost halved the risk of virologic failure at 18 months.

They conclude: “the impact of POC-VL monitoring on adherence and HIV transmission remains poorly understood despite these being the key factors that determine whether POC-VL will be cost-effective. To minimise unnecessary switches, the detection level should not be less than 1000 copies/mL which has important implications for the design of these tests.”