Abstract
Point-of-care (POC) HIV viral load (VL) testing offers the potential to reduce turnaround times for antiretroviral therapy monitoring, offer near-patient acute HIV diagnosis in adults, extend existing centralized VL services, screen women in labor, and prompt pediatrics to early treatment. The Liat HIV Quant plasma and whole-blood assays, prerelease version, were evaluated in South Africa. The precision, accuracy, linearity, and agreement of the Liat HIV Quant whole-blood and plasma assays were compared to those of reference technologies (Roche CAP CTMv2.0 and Abbott RealTime HIV-1) on an HIV verification plasma panel (n = 42) and HIV clinical specimens (n = 163). HIV Quant plasma assay showed good performance, with a 2.7% similarity coefficient of variation (CV) compared to the Abbott assay and a 1.8% similarity CV compared to the Roche test on the verification panel, and 100% specificity. HIV Quant plasma had substantial agreement (pc [concordance correlation] = 0.96) with Roche on clinical specimens and increased variability (pc = 0.73) in the range of <3.0 log copies/ml range with the HIV Quant whole-blood assay. HIV Quant plasma assay had good linearity (2.0 to 5.0 log copies/ml; R2 = 0.99). Clinical sensitivity at a viral load of 1,000 copies/ml of the HIV Quant plasma and whole-blood assays compared to that of the Roche assay (n = 94) was 100% (confidence interval [CI], 95.3% to 100%). The specificity of HIV Quant plasma was 88.2% (CI, 63.6% to 98.5%), and that for whole blood was 41.2% (CI, 18.4% to 67.1%). No virological failure (downward misclassification) was missed. Liat HIV Quant plasma assay can be interchanged with existing VL technology in South Africa. Liat HIV Quant whole-blood assay would be advantageous for POC early infant diagnosis at birth and adult adherence monitoring and needs to be evaluated further in this clinical context. LIAT cartridges currently require cold storage, but the technology is user-friendly and robust. Clinical cost and implementation modeling is required.
INTRODUCTION
HIV viral load (VL) testing is used to monitor the effectiveness of antiretroviral therapy (ART) after initiation, identify early virological failure and targeted adherence, and finally to provide guidance on when to switch therapy in late failures (1). VL testing has been used in the diagnosis of HIV acute infection (2) and early infant diagnosis (EID; <18 months of age) (3, 4); however,VL testing may be limited in the developing world, and EID still is widely performed with a qualitative test and VL is used as a secondary test (5). Several well-established VL technologies are available (6) in high-throughput and, more recently, ultrahigh-throughput platforms designed for centralized laboratories with good quality control. There is a dire need to increase access to VL testing in resource-limited settings (7) and to simplify and improve efficiency in diagnostics to ensure patient care is not compromised. Complexities related to logistics and sample transport to ensure RNA integrity in blood specimens is one limitation to providing full access to VL testing (8). Solutions to increasing access to VL testing being investigated are the use of dried blood spots (9–11), which are easy to transport and appear to extend sample RNA integrity (12), and on-site (point-of-care [POC]) VL testing (without requiring specimen transport and potential loss of specimen integrity) (13). POC VL testing, however, has been slow in commercialization, especially for large-scale implementation programs, such as that in South Africa. In spite of a rich diagnostic development pipeline and ample global support, including the Liat (laboratory-in-a-tube) analyzer (Roche Molecular Systems, NJ, USA), Alere q HIV-1/2 (Alere, Jena, Germany), SAMBA (Diagnostics for the Real World, Ltd., Cambridge, United Kingdom), GeneXpert viral load system (Cepheid, Sunnyvale, USA), and the EOSCAPE-HIV rapid RNA assay system (Wave 80 Biosciences, San Francisco, CA), among others (14, 15), few HIV VL POC tests have been evaluated and are commercially available (16, 17).
One such promising technology is the Liat HIV Quant POC VL assay (IQuum, Inc., Marlborough, MA, which was recently acquired by Roche Molecular Systems, Inc., Branchburg, NJ, USA) (18 and http://www.roche.com/media/media_releases/med-cor-2014-04-07.htm). A validation of this technology's initial plasma assay, requiring 200 μl plasma, showed good performance against the HIV-1 real-time test (Abbott Molecular, Des Plaines, IL, USA) (17). The current Liat HIV Quant molecular assay is a fully automated POC test that generates a quantitative HIV VL within 30 to 35 min using either 150 μl plasma (Liat HIV plasma Quant assay) or 75 μl whole blood (Liat HIV blood Quant assay). The ability to rapidly perform VL testing on both plasma and whole-blood specimens using the same instrument is an attractive feature, particularly where specialized skill is required for performing phlebotomy on infants, and it is not available for POC HIV VL testing in South Africa.
MATERIALS AND METHODS
Specimen collection and study site description.
Ethics approval was obtained from the University of the Witwatersrand, Johannesburg (M110139), for an additional 4 ml to 16 ml anticoagulated (EDTAK3) venous derived blood from adult patients attending two primary health clinics in Johannesburg for their routine antiretroviral therapy (ART) and tuberculosis (TB) monitoring between June 2012 and September 2012. The two clinics were Hillbrow Community Health Centre, Johannesburg (approval for 4 ml [1 tube] blood collection), and Themba Lethu HIV Clinic, Helen Joseph Hospital, Johannesburg (approval for 16 ml [4 tubes] blood collection). The blood specimens were couriered (∼30 min transport time) to the University of the Witwatersrand Research Diagnostics testing laboratory on the same day that venesection was performed. Whole-blood testing on the Liat HIV Quant assay was performed first, and then residual whole blood was centrifuged at 3,000 × g for 15 min using a Hettich EBA-20 centrifuge (Hettich AG, Germany), followed by plasma Liat HIV Quant assay testing. Residual plasma was tested on the Roche COBAS AmpliPrep/COBAS TaqMan (CAP/CTMv2) (Roche Molecular Systems, Inc., Branchburg, NJ, USA), and residual plasma with sufficient volume was tested with the Abbott RealTime HIV-1 assay (Abbott Diagnostics Abbott Molecular Inc., Des Plaines, IL, USA) using m2000sp and m2000RT. The Liat HIV whole-blood Quant assay was tested on the same day as blood draw, followed by plasma testing either the same or the next day (overnight storage at 4°C). Residual plasma was stored at −70°C, and batch testing on the comparator assays was performed in the accredited National Health Laboratory Service HIV PCR laboratory of the same department.
Liat HIV Quant assay VL testing.
Three Liat analyzers were couriered from the supplier (in Massachusetts, USA) to the University of the Witwatersrand in Johannesburg, where they were self-installed by the laboratory scientist using the manufacturer's package insert instructions. The Liat analyzer is a quantitative, fully automated instrument that performs silica magnetic bead sample extraction, multiplex real-time PCR amplification, and detection of HIV in a single assay tube and has a barcode reader and digital screen display with integrated keypad. The lower limit of detection is reported as 81 copies/ml in 150 μl plasma, and the dynamic range is 102 to 1.5 × 106 copies/ml. Both Liat HIV Quant assays (one assay tube for plasma and one assay tube for whole blood) have an internal control and require 75 μl whole blood or 150 μl plasma to generate results within 30 to 35 min. The testing cartridges require a cold chain (4°C).
Evaluation protocol and statistical methodology.
The blood specimens (whole blood and plasma) were tested on the Liat HIV Quant plasma and the Liat HIV Quant whole-blood assays to determine quantitative precision, accuracy (including misclassification), and linearity as well as the qualitative performance (ease of use, number of invalids and errors, and number of tests per day). An evaluation matrix is presented in Table 1 outlining the various evaluation components. A 42-member frozen plasma verification panel (termed the South African viral quality assessment [SAVQA], developed by the South African National Priority Program in Johannesburg using HIV subtype C plasma for verifying newly installed Roche CAP/CTMv2 and Abbott RealTime HIV-1 platforms), which included several confirmed HIV-negative blood specimens obtained from SANBS (South African Blood Transfusion Services) (19), also was included in the evaluation matrix. The panel is designed to measure precision (intra- and intervariability) as well as carryover (instrument contamination) and the limit of the blank (i.e., to correctly identify all HIV-negative specimens). Due to a limited number of Liat plasma cartridges available, 10 of 17 HIV-negative panel members were tested, and all (n = 25) of the quantifiable panel members were tested on a Liat analyzer. The results generated by the Liat HIV Quant plasma assay were compared to published criteria (<35% coefficient of variation on untransformed data; <0.19 log copies/ml standard deviations [SD]; <0.3 log copies/ml bias; <2.9% similarity coefficient of variation [CV] [21]) determined previously for the Roche CAP/CTMv2 and Abbott HIV-1 RealTime HIV-1 platforms on this panel (19).
TABLE 1.
Evaluation matrix of the specimen numbers, specimen type, and comparator VL tests performed
| Purpose | Material | No. and type of specimens | Comparator assay |
|---|---|---|---|
| Platform verification | Assessment-quality frozen plasma from 42-member panela | 10 HIV−, 25 HIV+ | Roche CAP/CTMv2; Abbott RealTime HIV-1 |
| Precision (intra- and intervariability) | 4 ml whole blood (and plasma) | 3 patient specimens (3, 4, and 5 log copies/ml) repeated 6 times; all tested on 3 instruments | Not applicable |
| Linearity and variability of LDL | Clinical specimen plasma diluted into HIV-negative plasma | 3 patient specimens (>5.0 log) serially diluted 1:10 down to 3.0 log copies/ml and then diluted 1:2 to 2.0 log copies/ml (repeated 9 times); all tested on 3 Liat HIV Quant assay instruments | Roche CAP/CTMv2 |
| Accuracy and misclassification | HIV+ clinical specimens | 157 clinical specimens tested on the CAP/CTMv2 and Liat HIV Quant assay plasma assays; 94 (of 157) tested by Liat HIV Quant assay whole blood and 63 (of 157) tested by RealTime HIV-1 | Roche CAP/CTMv2 |
Ten of 17 HIV-negative panel members were tested, and all (n = 25) of the quantifiable panel members were tested on a Liat analyzer due to limited numbers of cartridges being available at the time of study.
Liat HIV Quant plasma assay precision was determined for all three Liat analyzers and encompassed intra- and interinstrument precision and total precision [the square root of (intrainstrument SD2 + interinstrument SD2)]. The latter was established for both the whole blood and plasma Liat HIV Quant test cartridges for three VL log ranges (3.0 log copies/ml, 4.0 log copies/ml, and 5.0 log copies/ml). Intravariability was calculated using the means and SD across six replicate specimens per Liat analyzer. The intervariability was calculated from one (the same) specimen tested on each Liat analyzer.
To determine the Liat HIV Quant plasma assay linearity, three clinical plasma specimens with VL values of 5.0 log copies/ml were selected and serially diluted (1:10) in HIV-negative plasma down to 3.0 log copies/ml (Table 1). A further dilution of 1:2 was performed to obtain samples of 2.0 log copies/ml, and 9 repeats were prepared. Liat HIV Quant plasma assay and the reference testing (Roche CAP/CTMv2) were performed over 1 day on all instruments. Linearity was determined by linear regression (including R2).
Agreement (accuracy) was determined between the Liat HIV Quant plasma and the Roche CAP/CTMv2 and Abbott RealTime HIV-1 platforms, as well as between the Liat HIV Quant whole-blood assay and the Liat HIV Quant plasma assay on paired specimens. This was measured by concordance correlation (pc) (22), Bland-Altman difference plots (23), and percent similarity (24). Concordance correlation was calculated as a measure of agreement strength (accuracy and precision) between a new assay and a reference method (22, 25). The formula applied is pc = p × Cb (22, 25), where p is the Pearson correlation (a measure of precision) and Cb is the bias correction factor (a measure of accuracy). The value of pc (strength of agreement) is suggested to be <0.9 (poor), 0.90 to 0.95 (moderate), 0.95 to 0.99 (substantial), and >0.99 (almost perfect) (22). Any paired Liat HIV Quant whole-blood or plasma assay value of >1.0 log copies/ml difference from the reference was considered clinically relevant (i.e., the point at which patient management may change).
Clinical sensitivity, specificity, and misclassification.
The sensitivity and specificity (including the 95% confidence interval [CI]) at a clinical threshold of 1,000 copies/ml of the Liat HIV Quant plasma and whole-blood assays was determined using Roche CAP/CTMv2 as the reference. Detection by the Liat HIV Quant whole-blood and plasma assay quantifiable test results was compared to that of the Roche CAP/CTMv2 plasma VL as the reference method. The Roche CAP/CTMv2 results were sorted into 4 categories (1, TND [target not detected]; 2, lower detection limit [LDL; <1.59 log copies/ml]; 3, 2.0 to 2.99 log copies/ml; 4, 3.0 to >6.99 log copies/ml). Any paired Liat HIV Quant whole-blood or plasma assay value outside these reference categories were considered misclassified and were represented as a percentage (rate) of the number of reference test results in each category. Statistical analyses were performed using STATA version 12 (StataCorp, College Station, TX).
RESULTS
Performance of the Liat HIV Quant plasma assay using a standard verification panel.
The Liat HIV Quant plasma assay reported all HIV-negative and HIV-positive specimens correctly (100% concordance), with no carryover between specimen tests on the SAVQA (19) plasma panel. The Liat HIV Quant plasma assay had increased assay variability (reduced precision) in the low-VL ranges (≤3.0 log copies/ml). However, this increased variability was within the acceptable clinically relevant difference (bias) of 1.0 log copies/ml, and patient management would not have been changed. The Liat HIV Quant plasma assay demonstrated acceptable bias of <0.3 log copies/ml for the majority of VL values of >3.0 log copies/ml, with an overall acceptable percentage similarity CV of 2.7% compared to that of the Abbott RealTime HIV-1 assay and even better performance of 1.8% compared to that of Roche CAP/CTMv2. These are below the expected limit of 2.9% CV reported for the comparison between Roche CAP/CTMv2 and Abbott RealTime HIV-1 on the SAVQA panel (19). Twenty confirmed HIV-negative specimens all were reported as undetectable on both the Liat HIV Quant whole-blood and plasma assays.
Performance of the Liat HIV Quant plasma and whole-blood assay precision on clinical HIV-positive specimens.
The Liat HIV Quant whole-blood and plasma assay precision (variability) on clinical specimens shows overall good intravariability and intervariability, with median SD of <0.19 log copies/ml for whole blood and median SD of <0.22 log copies/ml for plasma across values of 3.0 log copies/ml to 5.0 log copies/ml. Increased variability, however, was noted with the Liat HIV Quant whole-blood assay in the category of 3.0 log copies/ml. The Liat HIV Quant plasma assay generated acceptable linearity (R2 = 0.99) for three analyzers across the range (5.0 log copies/ml to 2.0 log copies/ml). The variability at 2.0 log copies/ml for two analyzers (11% CV and 12% CV) was similar to that of the CAP/CTMv2 assay (12% CV), with one analyzer showing somewhat increased variability to 23% CV. The quantifiable results generated by the assays were the following: 94% (59/63) quantified by Abbott RealTime HIV-1, with a median of 4.2 log copies/ml; 93% (88/94) quantified by the Liat HIV Quant whole-blood assay, with a median of 4.5 log copies/ml; 88% (138/157) quantified by Roche CAP/CTMv2, with a median of 4.3 log copies/ml; and 82% (128/157) quantified by the Liat HIV Quant plasma assay, with a median of 4.3 log copies/ml. Of all HIV-positive clinical specimens analyzed by the Liat HIV Quant whole-blood, plasma, and comparator assays, the Liat HIV Quant plasma assay values cluster with those of Roche CAP/CTMv2, whereas the Abbott RealTime HIV-1 assay generates lower viral load values than both the Liat HIV Quant plasma assay and Roche CAP/CTMv2. Overall, the Liat HIV Quant assays align more with the Roche CAP/CTMv2 than Abbott RealTime HIV-1 assay (Fig. 1A and B).
FIG 1.
Modified Bland-Altman difference plots comparing the Liat HIV Quant plasma to the Roche CAP/CTM v2 reference comparator method (A) and the Abbott RealTime HIV-1 (HIV-1RT) reference comparator assay (B). Mean difference and SD lines are present, as well as an additional legend containing the concordance correlation, absolute bias, and percent similarity values for each comparison. The vertical axes are the paired differences, and the horizontal axes are the log copies/ml VL values for the reference technology.
The Liat HIV Quant whole-blood assay has more variability than the Liat HIV Quant plasma assay (Fig. 2), because it generates overall higher viral load values and has several outliers in the viral load range of <4.0 log copies/ml. Overall, the Liat HIV Quant whole-blood assay misclassified 16% and the Liat HIV Quant plasma assay misclassified 5% of samples compared to the reference Roche CAP/CTMv2 (Table 2). The misclassified specimens tested by the Liat HIV Quant plasma assay all had values of <3.0 log copies/ml (<1,000 copies/ml) and would not result in treatment switching. Testing on the Liat HIV Quant whole-blood assay, however, did misclassify 10 of 94 (10.6%) specimens that would have resulted in a change in patient management through upward misclassification. Neither the Liat HIV Quant plasma nor the Liat HIV Quant whole-blood assay generated any downward misclassification (i.e., virological failure [>1,000 copies/ml] classified by the reference) that would have been missed by either Liat HIV Quant assay. The sensitivity and specificity of the Liat HIV Quant plasma and Roche CAP/CTMv2 assays (n = 94) were 100% (CI, 95.3%, 100%) and 88.2% (CI, 63.6%, 98.5%), respectively. The sensitivity and specificity of the Liat HIV Quant whole-blood and Roche CAP/CTMv2 assays (n = 94) were 100% (CI, 95.3%, 100%) and 41.2% (CI, 18.4%, 67.1%), respectively.
FIG 2.
Modified Bland-Altman difference plot of the Liat HIV Quant plasma assay (considered the reference in this comparison) compared to the Liat HIV Quant whole-blood assay. The mean difference and SD lines are present, as well as an additional legend containing the concordance correlation, absolute bias, and percent similarity values for each comparison. The vertical axis is the paired difference, and the horizontal axis is the log copies/ml VL values for the reference technology.
TABLE 2.
Liat HIV Quant plasma and whole-blood assay detection of HIV compared to detection by Roche CAP/CTMv2 on HIV-positive clinical specimens
| Category of plasma VL (n = 94) | Result bya: |
||
|---|---|---|---|
| CAP/CTMv2 | Liat HIV Quant plasma assay | Liat HIV Quant whole-blood assay | |
| TND (log copies/ml) | TND | TND | 2.77 |
| TND | TND | 4.10 | |
| LDL (<1.60) (log copies/ml) | TND | TND | 4.83 |
| TND | TND | 3.29 | |
| 1.30 | TND | 3.70 | |
| 1.30 | TND | 4.08 | |
| 1.38 | TND | 3.38 | |
| 1.60 | 2.61 | 3.46 | |
| 1.60 | 2.37 | 4.17 | |
| 2.0–2.99 log (log copies/ml) | 2.17 | 1.86 | TND |
| 2.19 | TND | TND | |
| 2.30 | 2.41 | 3.83 | |
| 2.48 | 2.92 | TND | |
| 2.92 | 3.01 | 4.13 | |
| 2.99 | 3.03 | TND | |
| 3.0–6.99 log (n) | 79 | 79 | 79 |
| Overall misclassification (%) | NA | 5.3 (5/94) | 15.9 (15/94) |
| Misclassification resulting in a change in patient management (%) | NA | 0 | 10.6 (10/94) |
Underlined results indicate clinical relevance for ART monitoring (>1.0 log copies/ml difference at values of >1,000 copies/ml) compared to the reported comparator results (≤1,000 copies/ml).
Figure 1A and B further show that the assays with the most agreement between their paired viral load results are the Liat HIV Quant plasma assay and the Roche CAP/CTMv2. Their paired viral load values have substantial strength of agreement (pc = 0.96) with overall good accuracy and precision (5.4% CV) and acceptable variability in their bias (0.35 log copies/ml). The agreement between the Liat HIV Quant plasma and the Abbott RealTime HIV-1 is less (pc = 0.87) due to an increased variability in their bias (0.41 log copies/ml). The agreement between the Liat HIV Quant whole-blood assay and the Liat HIV Quant plasma assay (Fig. 2) shows all outliers occur in the plasma viral load range of <3.0 log copies/ml, with the Liat HIV Quant whole-blood assay generating higher viral load values with a concordance correlation of 0.81.
Qualitative analysis of the Liat HIV Quant assays.
The staff who performed this evaluation reported that the Liat analyzer installation was easy to perform, and no supplier support was necessary. This installation and self-training (familiarity with the assay features and testing procedures) were performed within 2 h, and no adverse events were experienced by the operators. The Liat analyzer had a small footprint (approximately 11.4 cm by 19 cm by 24.1 cm) with a touchscreen user interface. Throughout the testing period (n = ∼437 Liat HIV Quant assay tubes tested), seven errors were reported (1.6% error rate). These all were due to a scanning error; after rescanning, no specimen results were lost. A total of 12 to 14 specimens could be tested on a single Liat analyzer using either the HIV Quant whole-blood or plasma assay within a single 8-h day. As the assay cartridge is a closed system, no special safety precautions were required, and standard biohazardous waste disposal was needed. Data could be exported by USB port or Ethernet cable (however, no connectivity was used for this evaluation).
DISCUSSION
POC VL testing offers a solution to improved access, decreased result turnaround time, reduced patient return visits for results, immediate adherence counseling, and reduced concerns around sample integrity. The Liat HIV Quant assay has several features that are attractive to POC testing, such as a 30- to 35-min sample input to result time, ease of use by nonlaboratory personnel, and a small footprint. However, plasma testing would require the added step of centrifugation, and the current reagent has a short shelf life of 6 months and requires cold storage. The latter requirement, however, is not unachievable in areas where both assays and drugs use similar conditions in the clinic pharmacy. If used at the POC, however, viral load testing will require ongoing quality monitoring, which currently is accessible only to high-throughput laboratories and will require development for POC.
Overall, the Liat HIV Quant whole-blood and plasma assays show good precision and accuracy compared to existing technologies, such as the Roche CAP/CTMv2 and Abbott RealTime HIV-1. This was evident on the frozen plasma verification panel and fresh clinical specimens and also shows that the Liat HIV Quant whole-blood and plasma assays are well suited to HIV subtype C specimens representative of the South African HIV-infected population (although no sequencing was performed) (26). Overall, the agreement of the Liat HIV Quant assays are aligned more with that of Roche CAP/CTMv2 than that of Abbott RealTime HIV-1. The Liat HIV Quant assays and analyzer appear to have good reproducibility between and within repeat specimen testing for both whole blood and plasma and good linearity between 2.0 log copies/ml and 5.0 log copies/ml. However, increased variability was found in the viral load range of 2.0 log to 3.0 log copies/ml, which is similar to that for testing dried blood spots.
Compared to the Roche CAP/CTMv2, Liat HIV Quant plasma assay testing did not generate any clinically misclassified results that would have resulted in a change in patient management at the 1,000-copies/ml threshold. The Liat HIV Quant whole-blood assay testing, however, did yield 10.6% clinical misclassification on these same specimens, but these all were upward misclassified (i.e., identified higher viral load values than reference testing), and no specimens identified by reference testing as virological failure were missed. This is due to the assay's total nucleic acid extraction protocol, with amplification from whole-blood specimens of cell-associated HIV-1 DNA and RNA. This indicates repeat testing on plasma of all whole-blood specimens with Liat HIV Quant whole-blood values of ∼4.5 log copies/ml but is advantageous in assessing HIV-1 infection among newly infected persons receiving pre- or postexposure prophylaxis, where plasma HIV-1 RNA levels may be suppressed, and for newly infected neonates whose mothers had received ART at parturition. The sensitivity and specificity of the Liat HIV Quant whole-blood assay (100% [CI, 95.3%, 100%] and 41.2% [CI, 18.4%, 67.1%], respectively) compared to those of the Roche CAP/CTMv2 platform (n = 94) is not too dissimilar from values reported for dried-blood-spot VL monitoring (11); however, this study is limited by a small sample size.
Analysis of a national program's patient's HIV VL data has value in measuring a program's treatment performance (community VL) (27). However, this is possible only with central collection of VL results, such as through a laboratory information system (LIS). In South Africa, this is very well managed through the use of Abbott link, Roche Ampli-link (28), and Axeda systems; therefore, the introduction of new HIV VL technology must ensure connectivity in this context. The Liat analyzer does have full connectivity with HL7 communication protocols and wireless and Ethernet capabilities.
Positioning such an assay at POC in South Africa (currently performing ∼2 million VL analyses through centralized testing laboratories) would be best in underserviced areas and niche settings to extend access to VL testing. The goal of the South African national HIV treatment program was to add 500,000 new people to ART per year in their 3,540 clinics, which are staffed by ∼23,000 trained nurses (29). Areas where sample integrity may be compromised (>4-h specimen transport times), and where rapid turnaround of VL results is required (maternity wards to identify HIV-positive mothers and babies at birth), would benefit from using the Liat HIV Quant assays and ensure equity of access. The Liat HIV Quant plasma assay, as shown in this study, performs well against the two existing in-country VL plasma technologies, making it an ideal candidate for implementing in this existing program. Concerns would be minimal for cross-platform VL plasma test reporting, leading to clinical differences and unnecessary (or undetected) changes in patient management. Although the Liat HIV Quant assay was not evaluated on infant specimens in this study, whole-blood testing in the maternity setting would have the benefit of its total nucleic acid extraction protocol. However, both scenarios require further clinical evaluation (especially finger stick-derived whole-blood specimen testing), cost analyses (single-use specimens versus high throughout), procurement, and implementation modeling.
ACKNOWLEDGMENTS
We thank Matilda Nduna, Andrew Black, Francois Venter, Ian Sanne, and their clinic teams at WRHI and CHRU/RTC clinics in Johannesburg, Sergio Carmona and the NHLS PCR staff for performing the CAP/CTMv2 and HIV-1 RealTime HIV-1 comparator assays, and Shuqi Chen and his team at IQuum (Boston, MA) for loaning the instruments.
Grand Challenges Canada (grant 0007-02-01-01-01 [to W.S.]) also provided support.
REFERENCES
- 1.AIDSInfo. 13 November 2014, posting date Guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents: laboratory testing plasma HIV-1 RNA (Viral Load) and CD4 count monitoring. http://aidsinfo.nih.gov/guidelines/html/1/adult-and-adolescent-arv-guidelines/0/.
- 2.Chu C, Selwyn PA. 2010. Diagnosis and initial management of acute HIV infection. Am Fam Physician 81:1239–1244. [PubMed] [Google Scholar]
- 3.Delamare C, Burgard M, Mayaux MJ, Blanche S, Doussin A, Ivanoff S, Chaix ML, Khan C, Rouzioux C. 1997. HIV-1 RNA detection in plasma for the diagnosis of infection in neonates. The French Pediatric HIV Infection Study Group. J Acquir Immune Defic Syndr Hum Retrovirol 15:121–125. doi: 10.1097/00042560-199706010-00004. [DOI] [PubMed] [Google Scholar]
- 4.Read JS. 2007. Diagnosis of HIV-1 infection in children younger than 18 months in the United States. Pediatrics 120:e1547–e1562. doi: 10.1542/peds.2007-2951. [DOI] [PubMed] [Google Scholar]
- 5.AIDSInfo. 5 March 2015, posting date Guidelines for the use of antiretroviral agents in pediatric HIV infection: diagnosis of HIV infection in infants and children. http://aidsinfo.nih.gov/guidelines/html/2/pediatric-arv-guidelines/55/diagnosis-of-hiv-infection-in-infants-and-children.
- 6.Stevens WS, Scott LE, Crowe SM. 2010. Quantifying HIV for monitoring antiretroviral therapy in resource-poor settings. J Infect Dis 201(Suppl 1):S16–S26. doi: 10.1086/650392. [DOI] [PubMed] [Google Scholar]
- 7.Calmy A, Ford N, Hirschel B, Reynolds SJ, Lynen L, Goemaere E, Garcia de la Vega F, Perrin L, Rodriguez W. 2007. HIV viral load monitoring in resource-limited regions: optional or necessary? Clin Infect Dis 44:128–134. doi: 10.1086/510073. [DOI] [PubMed] [Google Scholar]
- 8.Bygrave H. 27 January 2014, posting date HIV viral load testing in Africa–no longer why but how? PLoS Blog http://blogs.plos.org/speakingofmedicine/2014/01/27/hiv-viral-load-testing-africa-longer/. [Google Scholar]
- 9.Johannessen A, Garrido C, Zahonero N, Sandvik L, Naman E, Kivuyo SL, Kasubi MJ, Gundersen SG, Bruun JN, de Mendoza C. 2009. Dried blood spots perform well in viral load monitoring of patients who receive antiretroviral treatment in rural Tanzania. Clin Infect Dis 49:976–981. doi: 10.1086/605502. [DOI] [PubMed] [Google Scholar]
- 10.Johannessen A, Troseid M, Calmy A. 2009. Dried blood spots can expand access to virological monitoring of HIV treatment in resource-limited settings. J Antimicrob Chemother 64:1126–1129. doi: 10.1093/jac/dkp353. [DOI] [PubMed] [Google Scholar]
- 11.Neogi U, Gupta S, Rodridges R, Sahoo PN, Rao SD, Rewari BB, Shastri S, Costa AD, Shet A. 2012. Dried blood spot HIV-1 RNA quantification: a useful tool for viral load monitoring among HIV-infected individuals in India. Indian J Med Res 136:956–962. [PMC free article] [PubMed] [Google Scholar]
- 12.Youngpairoj AS, Masciotra S, Garrido C, Zahonero N, de Mendoza C, Garcia-Lerma JG. 2008. HIV-1 drug resistance genotyping from dried blood spots stored for 1 year at 4 degrees C. J Antimicrob Chemother 61:1217–1220. doi: 10.1093/jac/dkn100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Schito M, Peter TF, Cavanaugh S, Piatek AS, Young GJ, Alexander H, Coggin W, Domingo GJ, Ellenberger D, Ermantraut E, Jani IV, Katamba A, Palamountain KM, Essajee S, Dowdy DW. 2012. Opportunities and challenges for cost-efficient implementation of new point-of-care diagnostics for HIV and tuberculosis. J Infect Dis 205(Suppl 2):S169–S180. doi: 10.1093/infdis/jis044. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.UNITAID. June 2013, posting date UNITAID HIV/AIDS diagnostic technology landscape, 3rd ed http://www.unitaid.eu/en/resources-2/news/9-uncategorised/345-technical-reports.
- 15.UNITAID. June 2014, posting date HIV/AIDS diagnostics technology landscape, 4th ed http://www.unitaid.eu/images/marketdynamics/publications/UNITAID-HIV_Diagnostic_Landscape-4th_edition.pdf.
- 16.Ritchie AV, Ushiro-Lumb I, Edemaga D, Joshi HA, De Ruiter A, Szumilin E, Jendrulek I, McGuire M, Goel N, Sharma PI, Allain JP, Lee HH. 2014. SAMBA HIV semi-quantitative test, a new point-of-care viral load monitoring assay for resource-limited settings. J Clin Microbiol 52:3377–3383. doi: 10.1128/JCM.00593-14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Coombs R, Dragavon J, Hargb S, Mangalindan O, Bremer J, Jennings C, Chen L, Tian Y, Chen S. 2011. Validation of a novel Lab-in-a-Tube analyzer and single-tube system for simple/rapid HIV-1 RNA quantification, poster 657 18th Conf Retrovir Opportun Infect, Boston, MA. [Google Scholar]
- 18.Tanriverdi S, Chen L, Chen S. 2010. A rapid and automated sample-to-result HIV load test for near-patient application. J Infect Dis 201(Suppl 1):S52–S58. doi: 10.1086/650387. [DOI] [PubMed] [Google Scholar]
- 19.Scott LE, Carmona S, Gous N, Horsfield P, Mackay M, Stevens W. 2012. Use of a prequalification panel for rapid scale-up of high-throughput HIV viral load testing. J Clin Microbiol 50:4083–4086. doi: 10.1128/JCM.01355-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Reference deleted.
- 21.Brambilla DJ, Granger S, Jennings C, Bremer JW. 2001. Multisite comparison of reproducibility and recovery from the standard and ultrasensitive Roche Amplicor HIV-1 Monitor assays. J Clin Microbiol 39:1121–1123. doi: 10.1128/JCM.39.3.1121-1123.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Lin LI-K. 2000. A note on the concordance correlation coefficient. Biometrics 56:324–325. doi: 10.1111/j.0006-341X.2000.00324.x. [DOI] [Google Scholar]
- 23.Bland JM, Altman DG. 1986. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet i:307–310. [PubMed] [Google Scholar]
- 24.Scott LE, Galpin JS, Glencross DK. 2003. Multiple method comparison: statistical model using percentage similarity. Cytom B Clin Cytom 54:46–53. doi: 10.1002/cyto.b.10016. [DOI] [PubMed] [Google Scholar]
- 25.McBride GB. 2005. A proposal for strength of agreement criteria for Lin.s.Concordance correlation coefficient. NIWA client report HAM2005-062. Report for Ministry of Health, Hamilton, New Zealand. [Google Scholar]
- 26.Papathanasopoulos MA, Cilliers T, Morris L, Mokili JL, Dowling W, Birx DL, McCutchan FE. 2002. Full-length genome analysis of HIV-1 subtype C utilizing CXCR4 and intersubtype recombinants isolated in South Africa. AIDS Res Hum Retrovir 18:879–886. doi: 10.1089/08892220260190362. [DOI] [PubMed] [Google Scholar]
- 27.Castel AD, Befus M, Willis S, Griffin A, West T, Hader S, Greenberg AE. 2012. Use of the community viral load as a population-based biomarker of HIV burden. AIDS 26:345–353. doi: 10.1097/QAD.0b013e32834de5fe. [DOI] [PubMed] [Google Scholar]
- 28.Kessler HH, Jungkind D, Stelzl E, Direnzo S, Vellimedu SK, Pierer K, Santner B, Marth E. 1999. Evaluation of AMPLILINK software for the COBAS AMPLICOR system. J Clin Microbiol 37:436–437. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.McNeil DG. 25 August 2014, posting date AIDS progress in South Africa is in peril. The New York Times. http://www.nytimes.com/2014/08/26/health/aids-south-africa-success-pepfar.html?_r=1. [Google Scholar]