Apparent low-level HIV RNA viraemia related to sample processing time MD Portman 1 and CJ Lacey 2 1 Genitourinary Medicine, Leeds General Infirmary, Centre for Sexual Health, Leeds, West Yorkshire, UK and 2 Centre for Immunology and Infection, University of York, York, North Yorkshire, UK We are concerned regarding Chapter 11 of the draft British HIV Association (BHIVA) monitoring guidelines ‘Technical aspects of viral load testing’ [1]. This states that ‘based on available information, viral RNA in blood samples col- lected into EDTA tubes is stable for at least 2–3 days at room temperature, allowing transportation of the sample by post or collection over a weekend’. We believe that the references cited [2,3] may not be applicable to current practice because they relate to the stability of HIV-1 RNA in whole blood in patients who, crucially, are not taking antiretroviral therapy (ART). There is current concern regarding low-level viraemia in patients on ART [4] which is incompletely understood. We believe that time to processing of samples for HIV-1 RNA testing plays an important part in the genesis of low-level viraemia. At our HIV clinic in York we observed more patients on ART with detectable viral loads than expected and there- fore conducted a service evaluation during March to May 2009. We took paired samples for HIV-1 RNA testing from 21 patients who had been stable on ART for 6 months. One sample had plasma separated in York Microbiology Department (York Teaching Hospital NHS Foundation Trust) prior to transportation to the virology lab in Leeds and the other was transported as whole blood in an ethylenediaminetetraacetic acid (EDTA) monovette tube. The mean time to local centrifugation was 4 hours and to processing at the virology lab was 28 hours. Samples were assayed using the Roche TaqMan v2.0 (COBAS® AmpliPrep/COBAS® TaqMan® HIV-1 Test, Roche Molecular Diagnostics, UK). We found that nine of 21 whole-blood samples (43%) had an HIV-1 viral load above 400 HIV-1 RNA copies/mL, i.e. at a level where resistance testing or therapeutic drug monitoring would be instigated and treatment augmentation/switch considered [5]. In contrast, no sepa- rated sample had a viral load > 400 copies/mL. Twelve of 21 whole-blood samples (57%) had an HIV-1 viral load > 200 copies/mL, i.e. at a level whereby patients were recalled for discussion on adherence and resistance testing considered. No separated sample had a viral load > 200 copies/mL. Only two whole-blood samples had a viral load of < 40 copies/mL compared with 19 of 21 separated samples (90%). All separated samples had an HIV-1 viral load of 54 copies/mL or less, i.e. nil had a significant viraemia (Fig. 1). The range of results for whole-blood samples was from ‘not detected’ to 3080 copies/mL; the mean was 629 copies/mL and the median 279 copies/mL. Further research in this important area is needed. HIV-1 RNA results that are above the cut-off in patients on treatment have much greater implications than a slight- ly inaccurate result in a patient off treatment. There is currently no evidence in the literature which relates to the reproducibility of HIV RNA assays at low copy number relating to different periods of time pre- centrifugation in patients on ART. Therefore, until these Correspondence: Dr Mags D. Portman, Genitourinary Medicine, Leeds General Infirmary, Centre for Sexual Health, Sunnybank Wing, Great George Street, Leeds, West Yorkshire, LS1 3EX, UK. E-mail: mags.portman@gmail.com 0 2 4 6 8 10 12 14 16 18 20 <40 40-199 200- 399 >400 HIV VL (copies/ml) Number of samples Separated at 28 h Separated at 4 h Fig. 1 Difference between HIV viral loads (VLs) in samples separated at mean 4 and 28 h. DOI: 10.1111/j.1468-1293.2012.01016.x © 2012 British HIV Association HIV Medicine (2012), 13, 578–579 LETTER TO THE EDITOR 578