Field study to determine the reliability of HIV viral load results shows minimal impact of delayed testing in South Africa

Abstract

Background: Understanding factors that impact HIV viral load (VL) accuracy in resource-limited settings is key to quality improvement.

Objective: We evaluated whether testing delay and specimen storage between 25 °C and 30 °C before testing affected results.

Methods: Between November 2019 and June 2023, 249 individuals on antiretroviral therapy, or with newly diagnosed HIV, were recruited from clinics in Cape Town and Gqeberha, South Africa, and three plasma preparation tubes were collected. One tube was tested within 24 h, while the others were stored uncentrifuged at ambient temperatures before testing. Centrifugation and testing of matched samples were performed on Day 4 and Day 7 after collection.

Results: Time delay and ambient storage had minimal impact in specimens with a Day 1 VL of > 100 copies/mL. When grouped by Day 1 VL range, 96% - 100% of specimens at Day 4 and 93% - 100% at Day 7 had VLs within 0.5 log copies/mL of the first result. The greatest variability at Days 4 and 7 was observed when the Day 1 VL was < 100 copies/mL. However, there was no trend of increasing difference over time. Of Day 1 specimens with undetectable VL, or VL < 50 copies/mL, 80% had concordant results at Day 4 and 78% at Day 7.

Conclusion: These results show that VL is stable in plasma preparation tubes for 7 days when stored at room temperature. There is significant variability in specimens with low VL, but variability is not affected by testing delay.

What this study adds: Ideal HIV VL testing conditions are frequently unachievable in resource-limited settings. Data are needed on whether this impacts on the validity of test results. Our results provide reassurance that storage at ambient temperature for up to 7 days before testing does not substantially affect the VL result.

Keywords: HIV viral load stability; delayed testing; diagnostic accuracy; high-throughput viral load testing; plasma preparation tubes.

FIGURE 1

Study plan for testing factors affecting accuracy of viral load samples, South Africa, 2019–2023.

FIGURE 2

Variability of all paired viral load results at Days 4 and 7, South Africa, 2019–2023. Bland-Altman plots of Day 1 versus Day 4 VL (a) and Day 1 versus Day 7 (b) show that viral load measurements tended to be systematically higher by an average of 0.12 log at Day 4 and 0.14 log at Day 7 as compared to Day 1. The bias tended to be greater in samples with low RNA levels (< 2 log copies/mL); (c) shows variability of paired viral loads of Day 1 samples that were re-run (R) on Day 7, with the Bland Altman plot revealing minimal bias and even lower variance in Day 1 versus Day 7 results. Green lines represent +2 and −2 standard deviations and the red dotted line represents the mean difference in HIV viral load on the Bland Altman plots.