Six-monthly viral load testing of patients taking antiretroviral therapy (ART) at primary health clinics in rural Kenya reduced the risk of virologic failure at 18 months of follow-up by 46%, according to the results of the Clinic-based ART Diagnostic Evaluation (CLADE) randomised controlled trial conducted in rural Kenya at seven district-level clinics. Dr Frederick Sawe of the Kenya Medical Research Institute presented the findings to the 20th Conference on Retroviruses and Opportunistic Infections (CROI 2013) in Atlanta last week.
Routine viral load monitoring is still the exception rather than the rule in resource-limited settings owing to a number of barriers to implementation. The lack of point-of-care viral load testing, coupled with the need to send blood samples to laboratories that can carry out the test, make it difficult to carry out viral load testing for every patient.
WHO guidelines recommend the use of HIV-1 RNA viral load to confirm ART treatment failure, and routine viral load tests every six months where available. Kenya Ministry of Health (MoH) guidelines recommend targeted viral load if treatment failure is suspected. However, the WHO guidance is based on “low quality evidence”. Randomised controlled trial data, including cost effectiveness in rural settings where the majority of the HIV burden exists, are lacking.
The CLADE study randomised 820 treatment-naive patients to receive standard-of-care clinical and immunologic monitoring with confirmatory or targeted viral load (Routine Care Arm), or clinical and immunological monitoring with concurrent viral load at every six months (Viral Load Arm). Participants were followed up for 18 months.
The study took place in seven rural district clinics, five of which are managed by the Kenyan Ministry of Health and two of which managed by faith-based organisations. The clinics are primarily staffed by clinical officers and nurses. The baseline characteristics were similar between the two groups, with 57% of participants being female, 37 years being the median age and the median CD4 count being between 164 and 168 cells/mm3 in both groups. Just under one-third had advanced HIV disease (WHO stages 3 or 4 HIV disease).
Viral load testing was carried out at a central laboratory in Kericho.
Viral failure (HIV-1 RNA greater than 1000 copies/ml)in the Routine Care Arm was 15.2% (51 of 336) compared to 8.7% (28 9f 321) in the Viral Load Arm (p = 0.006). Overall, 12% of all study participants (79 of 657) experienced viral failure. Multivariate analysis showed that the 18-month risk of viral failure in the viral load monitoring arm was 0.54 (0.35 – 0.87) when compared to the standard-of-care arm.
Of the 820 patients enrolled in the study, 34 (4.1%) had treatment failure and began second-line ART. Four (0.9%) of these were in the Routine Care Arm and 30 (7.3%) were in the Viral Load Arm (p<0.01). Overall, 336 (82.2%) in the Routine Care Arm and 321 (78.5%) in the Viral Load Arm completed the per-protocol, 18-month final visit within the 21-day window.
The study also found that 15.4 people needed to be screened in order to prevent one viral failure.
Overcoming barriers to viral load testing
Several strategies to improve access to viral load testing are being pursued simultaneously. As well as the development of point-of-care testing devices, which are still some way off, research groups have been investigating interim measures that could make it easier to ship blood samples and bring down the cost of laboratory-based testing. A number of research groups presented data at the conference on innovative methods of sample collection and testing for the monitoring of viral failure.
Three groups presented data on the use of dried blood spots as a medium of collection for blood samples. Dried blood spots have a number of potential limitations, including concerns about stability at high temperatures and over time, and questions about the correlation between HIV RNA levels in plasma and whole blood, which might lead to poorer sensitivity of dried-blood-spot testing.
A large ANRS study which compared the performance of dried-blood-spot and plasma samples for the detection of virologic failure (above 1000 copies/ml) showed that dried-blood-spot sampling varied in sensitivity from 84 to 93% according to the viral load assay used (Abbott m2000rt had the highest sensitivity score (Ayouba). A South African study found that 15% of dried-blood-spot samples produced results that miscategorised patients as either virologically suppressed or virologically failing, but that duration of time between sampling and testing did not appear to influence the accuracy of the result (Wallis). MSF researchers in Malawi showed that fingerprick blood sampling for dried blood spots was sufficient to generate samples appropriate for viral load testing. Testing of dried-blood-spot samples had a sensitivity of 90% when compared with plasma sampling (Metcalf).
MSF researchers in Mozambique evaluated a point of care nucleic acid testing (NAT) platform used for infant HIV DNA testing as a means to identify virologic failure in adults receiving antiretroviral therapy (Alere point-of-care NAT), again using fingerprick blood sampling in comparison to venous samples. The NAT device had a sensitivity of 95% for detecting virologic rebound above 5000 copies/ml, and a sensitivity of 98.5% for early infant diagnosis (Jani). Pooled nucleic acid testing, as used in testing of blood donations, was evaluated in South Korea. 215 samples were tested in pools of five and the method was found to be 100% accurate in detecting virologic failure, while saving $15,912 in viral load testing costs during 10 months (Kim).
Pooled viral load testing was also evaluated by MSF in Malawi, once again in pools of five, and found to have a negative predictive value of 98.9% at a viral load cut-off of 1000 copies/ml. Three hundred and fifty samples were tested; researchers estimated that pooled testing would save between $2402 and $4318 on this batch of tests alone (Pannus).
Ayouba A et al. Dried blood spots for HIV-1 viral load and drug resistance monitoring in HAART-treated patients from Africa and Asia: The Agence de Nationale de Recherche sur le Sida 12235 study. 20th Conference on Retroviruses and Opportunistic Infections, Atlanta, abstract 606, 2013.
Jani I et al. Evaluation of point-of-care nucleic acid testing for HIV viral load and early infant diagnosis in primary health clinics: Mozambique. 20th Conference on Retroviruses and Opportunistic Infections, Atlanta, abstract 607, 2013
Kim H et al. Pooled nucleic acid testing to identify ART failure during HIV infection in a tertiary teaching hospital: Seoul, South Korea. 20th Conference on Retroviruses and Opportunistic Infections, Atlanta, abstract 611, 2013
Metcalf C et al. Use of finger-prick dried blood spots for quantifying HIV-1 viral load, a diagnostic accuracy study: Thyolo, Malawi. 20th Conference on Retroviruses and Opportunistic Infections, Atlanta, abstract 608, 2013.
Sawe, F et al. Superiority of routine viral load monitoring in rural Kenya: the Kericho Clinic-based ART Diagnostic Evaluation (CLADE) Trial. 20th Conference on Retroviruses and Opportunistic Infections, Atlanta, abstract 151, 2013
Wallis C et al. Refining approaches to viral load monitoring using dried blood spots. 20th Conference on Retroviruses and Opportunistic Infections, Atlanta, abstract 604, 2013.