Waters, Laura MRCP*; Kambugu, Andrew MD, MMed†; Tibenderana, Hilda MD†; Meya, David MD†; John, Laurence MRCP*†; Mandalia, Sundhiya MSc*; Nabankema, Maggie†; Namugga, Irene†; Quinn, Thomas C MD‡§; Gazzard, Brian FRCP*; Reynolds, Steven J MD‡§; Nelson, Mark MD*
From the *St. Stephens AIDS Trust, Chelsea & Westminster Hospital, London, UK; †Infectious Disease Institute, Makerere University, Kampala, Uganda; ‡National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD; and the §Johns Hopkins University School of Medicine, Baltimore, MD.
Received for publication February 6, 2007; accepted August 29, 2007.
Funded by grants from the St. Stephen's AIDS Trust and the US National Institute of Allergy and Infectious Diseases, National Institutes of Health International Centers for Excellence in Research Program.
Data presented at the Annual Conference of the British HIV Association, Glasgow, April 25-28, 2007 (oral abstract 23).
Correspondence to: Laura Waters, MRCP, St. Mary's Hospital, London W2 1NY, UK (e-mail: firstname.lastname@example.org).
Background: Most HIV-infected subjects on antiretroviral therapy (ART) in resource-limited settings do not undergo virologic monitoring. There is an urgent need for cheap, accessible HIV RNA assays for early diagnosis of virologic failure. We investigated filter paper transfer (FPT) of whole blood and plasma as an alternative to standard plasma-based assays for virologic monitoring in Uganda.
Methods: Whole blood (n = 306) and plasma (n = 218) from 402 subjects established on ART were spotted onto filter paper and transported to Europe for HIV RNA extraction and quantification. These results were compared to a gold standard plasma assay in Kampala.
Results: Of 402 ART-treated subjects, 39 (9.7%) had viremia detectable (>500 copies/mL) by local methods. Plasma FPT showed excellent agreement with gold standard, whereas whole blood yielded a large number of false-positive viral loads.
Conclusions: This is the first study to investigate the use of FPT in ART-treated subjects and demonstrates that it may provide a practical, reliable method for virologic monitoring in resource-poor settings. Plasma FPT was accurate but requires centrifuge; whole blood produced a high number of false-positive results, but these were low-level. Whole blood may be sufficiently accurate if higher HIV RNA cut-offs were used to define virologic failure.
Over 1 million HIV-infected individuals are now established on antiretroviral therapy (ART) in Africa.1 The vast majority of these treated patients do not undergo routine HIV RNA quantification due to the cost and complexity of providing reliable assays in resource-limited settings.2 Filter paper transfer (FPT) of dried whole-blood spots to central laboratories is currently being used in Africa for the diagnosis of neonatal HIV (HIV-DNA).3 FPT of whole-blood and plasma spots has demonstrated reliability for HIV RNA quantification (viral load measurement) in untreated HIV-infected adults in both the developed world and Africa.4,5 There are no published data describing the performance of FPT, of whole blood or plasma, for HIV RNA quantification in ART-treated subjects. If shown to be reliable for this purpose, then FPT could enormously simplify virologic monitoring, and potentially reduce the associated costs, in Africa and other resource-limited settings.
In this study, blood samples obtained from 402 individuals established on ART in Uganda underwent viral load testing at the Infectious Diseases Institute clinic in Kampala (gold standard) and FPT testing. Local viral load testing was performed on liquid plasma using the Roche standard reverse transcriptase polymerase chain reaction (RT-PCR) based assay (Amplicor HIV-1 Monitor v1.5; Roche Molecular Systems, Belleville, NJ). Whole blood and/or plasma from the same patient samples were spotted onto filter paper (Whatman 903; Maidstone, UK), air-dried, and stored in sealed envelopes at ambient temperature. Filter paper samples were then sent to Europe every 3 weeks for HIV RNA extraction (Primagen, Amsterdam, The Netherlands) and quantification using the COBAS TaqMan real-time reverse transcriptase PCR assay (Roche Molecular Systems) with a lower limit of detection of 500 copies/mL. Filter paper samples were sent to Europe because the filter paper elution and HIV RNA extraction process was not available in Uganda at the time of this study. Initially whole blood was collected, but after interim analysis, FPT of plasma and whole blood was performed, followed by plasma samples only. FPT results were compared with gold standard results in terms of correct undetectable (<500 copies/mL) and detectable (>500 copies/mL) viral loads. Because an unplanned analysis was performed during the initial phase of the study, the whole-blood data were analyzed at a greater significance level (99%). Linear regression analyses was used on log10-transformed data to establish the equation of the trend between the gold standard and FPT results. The results are presented as the proportion of variance (explained by the scatter of the plot), together with the correlation coefficient, which indicates strength of the association between the gold standard and FPT results.
In total, 524 FPT viral loads were performed in 402 ART-treated patients. All subjects were receiving first-line therapy of a nonnucleoside reverse transcriptase inhibitor plus 2 nucleoside analogues. A total of 63% received nevirapine/stavudine/lamivudine and 37% efavirenz/zidovudine/lamivudine; median duration of therapy was 11 months. The FPT tests performed are detailed in Table 1.
Thirty-nine individuals (9.7%) undergoing HIV RNA quantification had a detectable viral load by local testing (median, 15,161 copies/mL; range, 511 to 447,000 copies/mL). Of the 524 filter paper samples, 306 were whole blood and 218 were plasma (including 122 whole-blood/plasma pairs). All filter paper samples were processed within 4 weeks of venesection. The positive (>500 copies/mL) and negative (≤500 copies/mL) HIV RNA results for both plasma and whole-blood FPT, compared with gold standard, are illustrated in Table 2A and B, respectively.
Compared with gold standard results, whole-blood filter paper specimens yielded 4 false-negative (all <2000 copies/mL by gold standard testing) and 64 false-positive results (median, 1002 copies/mL; range, 510 to 3510 copies/mL). Whole-blood FPT therefore had a sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of 0.86 (0.67 to 0.96), 0.77 (0.71 to 0.81), 0.27 (0.18 to 0.46), and 0.98 (0.95 to 1.00), respectively (99% confidence limits; Table 3). The McNemar test showed a significant difference between the gold standard test and the whole-blood FPT (P < 0.001).
Plasma FPT yielded no false-negative and 1 false-positive result (593 copies/mL). The sensitivity, specificity, and positive and negative predictive values of plasma FPT were therefore 1.00 (0.84 to 1.00), 0.99 (0.97 to 1.00), 0.95 (0.77 to 1.00), and 1.00 (0.77 to 1.00), respectively (95% confidence limits; Table 3).
The McNemar test showed no significant difference between the gold standard and plasma FPT viral loads (P = 0.317). Bland-Altman plots confirm the high number of low viral load false positives for whole-blood FPT but demonstrate good agreement (within 2 standard deviation [SD]) between whole-blood and plasma FPT for higher viral loads (>5000 copies/mL). Finally, the 122 whole-blood/plasma pairs were analyzed for agreement using the Cohen κ statistic. The observed agreement was fair at 73% (64% to 80%); z = 3.89 (P < 0.001).
Whole-blood and plasma FPT viral loads correlated significantly with the gold standard (Figs. 1 and 2). Figure 1 illustrates the numerous false positives with whole-blood FPT.
This study has demonstrated that FPT of plasma specimens for HIV RNA quantification may provide a practical and reliable means of monitoring ART in resource-poor settings. Filter paper samples could be collected from peripheral clinics and transported at tropical ambient temperatures within 4 weeks to central laboratories with the technology to perform sample elution and RNA extraction from FPT samples and HIV RNA RT-PCR. This would avoid the need to provide and sustain laboratories capable of performing HIV RNA quantification at each clinic site and might facilitate a reduction in the cost per viral load test by enabling the bulk purchase of assay kits. Unfortunately, whole-blood samples, which may be obtained by a simple finger-prick spot directly onto filter paper, were associated with a high number of false-positive detectable viral loads in our study; these could lead to switching therapy unnecessarily, using up valuable drug options. A cross-reaction between cell-associated HIV DNA and the HIV RT-PCR RNA assay is the likely explanation for the high rate of false-positive results. Plasma samples, although more accurate in our study, require venesection, blood bottles containing ethylenediamine tetraacetic acid (EDTA), and electricity-dependent centrifuge equipment, which would be a problem for many peripheral clinics. Possible solutions include using manual or battery-operated centrifuges, obtaining plasma from EDTA samples left to stand for 6 hours, or using nucleic acid sequence-based amplification (NASBA) assays that require HIV DNA detection to quantify HIV RNA and that therefore may be used on whole blood.6
Because second-line options in resource-poor settings are scarce, the use of a higher HIV RNA cut-off may be more appropriate when making decisions to switch therapy.7 If, for example, an HIV RNA load of 10,000 copies/mL was deemed an appropriate level at which to change regimen,7 whole-blood DBS may provide sufficient accuracy for this strategy, because in our study all false-positive whole-blood FPT results were below 5000 copies/mL and all false-negative results were on samples with a viral load <2000 copies/mL. The optimal HIV RNA cut-off for switching patients in resource-limited settings require investigation.
Finally, because this is the first study describing the use of FPT techniques for monitoring ART-treated individuals, further work to validate our results is required, ideally at multiple centers. In addition, data for different HIV subtypes and from samples collected under different field conditions may be useful. Optimal methods for sample elution and RNA extraction from filter paper specimens need to be clarified and the methodology rolled out to central laboratories in Africa and other resource-limited settings.
The authors thank Esther de Rooij, Robert Colebunders, Dave Thomas, Oliver Laeyendecker, Paul Walsh, Adrian Wildfire, Keith McAdam, and the Academic Alliance.
2. Crowe S, Turnbull S, Oelrichs R, et al. Monitoring of human immunodeficiency virus infection in resource-constrained countries. Clin Infect Dis. 2003;37(Suppl 1):S25-S35.
3. Panteleeff DD, John G, Nduati R, et al. Rapid method for screening dried blood samples on filter paper for human immunodeficiency virus type 1 DNA. J Clin Microbiol. 1999;37:350-353.
4. Brambilla D, Jennings C, Aldrovandi G, et al. Multicenter evaluation of use of dried blood and plasma spot specimens in quantitative assays for human immunodeficiency virus RNA: measurement, precision, and RNA stability. J Clin Microbiol. 2003;41:1888-1893.
5. Mwaba P, Cassol S, Nunn A, et al. Whole blood versus plasma spots for measurement of HIV-1 viral load in HIV-infected African patients. Lancet. 2003;362:2067-2068.
6. Holodniy M. HIV-1 load quantitation: a 17-year perspective. J Infect Dis. 2006;194(Suppl 1):S38-S44.
7. World Health Organization. Antiretroviral therapy for adults and adolescents in resource limited settings: towards universal access (2006 revision). Available at: http://www.who.int/hiv/pub/guidelines/adult/en/
. Accessed July 7, 2007.
© 2007 Lippincott Williams & Wilkins, Inc.