Abstract
HIV-1 drug resistance (HIVDR) assays are important tools in clinical management of HIV-infected patients on antiretroviral therapy (ART) and surveillance of drug-resistant variants at population levels. The high cost associated with commercial assays hinders their use in resource-limited settings. We adopted and validated a low-cost in-house assay using 68 matched plasma and dried blood spot (DBS) samples with a median viral load (VL) of 58,187 copies/ml, ranging from 253 to 3,264,850 against the commercial assay ViroSeq. Results indicated that the in-house assay not only had a higher plasma genotyping rate than did ViroSeq (94% versus 78%) but also was able to genotype 89.5% (51/57) of the matched DBS samples with VLs of ≥1,000 copies/ml. The sensitivity in detecting DR mutations by the in-house assay was 98.29% (95% confidence interval [CI], 97.86 to 98.72) on plasma and 96.54 (95% CI, 95.93 to 97.15) on DBS, and the specificity was 99.97% (95% CI, 99.91 to 100.00) for both sample types compared to ViroSeq. The minor DR mutation differences detected by the in-house assay against ViroSeq did not result in clinical significance. In addition, cost analysis showed that the in-house assay could reduce the genotyping cost by about 60% for both plasma and DBS compared to ViroSeq. This field condition evaluation highlights the potential utility of a cost-effective, subtype-independent, in-house genotyping assay using both plasma and DBS specimens for HIVDR clinical monitoring and population-based surveillance in resource-limited settings.
INTRODUCTION
There has been a tremendous increase in antiretroviral therapy (ART) in sub-Saharan Africa and other developing countries, largely due to the increased support by various multinational groups, such as the U.S. President's Emergency Plan for AIDS Relief (PEPFAR) and the Global Fund to Fight AIDS, Tuberculosis and Malaria (1–4). This has resulted in the significant reduction of HIV/AIDS-related morbidity and mortality among the 1.6 million patients currently on ART in the sub-Saharan region (1, 4). The long-term success of these ART programs, however, requires adequate monitoring of ART patients to ensure favorable treatment outcomes and minimize the development and transmission of HIV drug resistance (HIVDR), given the limited antiretroviral (ARV) drug regimen choices available in these settings (5–7). The cost and logistics involved in assays that are used to monitor HIVDR in ART patients remain challenging in resource-limited settings (5, 8, 9). Several initiatives are under way to establish alternative, less costly methods for laboratory-based ART patient monitoring, such as point-of-care CD4 testing, semiquantitative HIV viral load (VL) testing, and HIVDR testing using dried blood spots (DBS) (10–16). These parameters are routinely used in assessing treatment responses in resource-rich countries. HIVDR tests for patients on ART are not only important in monitoring individual patient treatment outcomes but also essential as public health tools in the routine assessment of the spread of drug-resistant variants at population levels (5, 17–19). Such data are vital in guiding a country's ARV drug regimen implementation strategy and in forecasting the need for new ARV drugs, especially in resource-limited settings where treatment options are limited (20–22).
Currently, there are two HIV-1 genotyping systems approved by the U.S. Food and Drug Administration (FDA), ViroSeq (Abbott Molecular, Abbott Park, IL) and TruGene (Siemens Healthcare Diagnostics, Deerfield, IL). These commercial systems are costly and were designed and approved to genotype HIV-1 subtype B viruses, which makes them less sensitive in genotyping HIV-1 variants in geographic areas where non-B subtypes and circulating recombinant forms (CRFs) are predominant (23, 24).
The preferred sample type used in these assays is either plasma or serum. This type of sample requires cold-chain conditions for collection, transportation, and storage (23) as well as separation of plasma from whole blood within 6 h of blood collection. In addition, venous blood collection requires well-trained phlebotomists and poses a potential risk to health care workers for occupational HIV exposure from needle sticks. In resource-limited settings, inadequate health care infrastructure often necessitates the collection of blood samples from peripheral sites and transport to a reference facility where processing of samples and testing are carried out. These logistical issues make the conventional plasma specimen collection nonideal for use in such settings. DBS offer an alternative specimen type to overcome these challenges: DBS do not require large amounts of blood and can be collected easily by finger or heel prick with minimal preprocessing procedures or risk to the health care worker. Moreover, they are easier to transport and store, do not require cold-chain transportation, and can be stored at ambient temperature for up to 2 weeks without compromising the genotyping efficiency (25, 26). Once air dried and properly packaged, they are considered noninfectious (27).
In order to mitigate the high cost associated with commercial genotyping tests and streamline the sample collection process, there have been systematic efforts to develop and evaluate in-house assays and adopt the assays for use with the DBS sample type, and several assays have shown promising results (24, 28–30). With the publication of the WHO manual for HIV drug resistance testing using dried blood spot specimens in 2010 and the promising results obtained from the in-house genotyping assays (31), we initiated a study to adopt and validate a broadly sensitive in-house assay (32, 33) against the FDA-approved ViroSeq HIV-1 drug resistance genotyping system using both plasma and DBS specimens.
MATERIALS AND METHODS
Study population.
Plasma and DBS samples were obtained from HIV-1-positive mothers and children enrolled in the Kisumu Breastfeeding Study (KiBS), which assessed the efficacy of combined ART, mainly either nevirapine (NVP) with zidovudine (AZT) and lamivudine (3TC) or nelfinavir (NFV) with AZT and 3TC given from 34 weeks into the gestation period through 6 months postpartum for the prevention of mother-to-child transmission (PMTCT) (34). These included 39 samples from mothers collected at 6 months postpartum after at least 6 months on ART and 29 samples from infants exposed to maternal ART through breast milk. The VL of these specimens ranged from 253 to 3,264,850 copies/ml. We also included seven DBS samples from a proficiency panel which were used to assess the reproducibility of DBS. These were replicates from a 14-member panel from the virological quality assurance program (VQA) at Chicago, IL, that had been contracted by the WHO-sponsored HIV proficiency testing program (30).
DBS and plasma collection and storage.
Plasma was harvested through separation within 6 h from the time of whole-blood collection in EDTA-treated anticoagulant Vacutainer tubes (Becton, Dickinson, San Jose, CA) and stored at −60°C to −80°C. DBS were prepared by pipetting 50 μl of whole blood onto each of 5 spots on a 903 Whatman filter paper card (Schleicher & Schuell, Keene, NH). Filter papers were dried overnight in a clean biosafety cabinet and then placed individually in an air-impermeable zip-lock bag containing desiccants and a humidity indicator and stored at −30°C.
Nucleic acid extraction from plasma and dried blood spots.
RNA from plasma was extracted using the FDA-approved ViroSeq HIV-1 genotyping system (Abbott Molecular, Abbott Park, IL) extraction protocol and the QIAamp viral RNA minikit (Qiagen Inc., Chatsworth, CA) for the in-house HIV-1 genotyping assay according to the manufacturer's instructions. Total nucleic acid (TNA) from DBS was extracted using a modified NucliSENS silica-based boom method (bioMérieux, Inc., Durham, NC) (29). Briefly, two 6-mm spots from each patient sample were cut and added into a tube containing 0.9 ml NucliSENS lysis buffer (bioMérieux, Inc., Durham, NC) and incubated at room temperature for 2 h under gentle rotation. After incubation, the tube was centrifuged for 2 min at 1,500 × g, and the supernatant was transferred into new 2-ml tubes. Nucleic acid was then extracted according to the manufacturer's instructions, eluted using 60 μl of elution buffer, and stored at −80°C until use.
HIV-1 drug resistance genotyping.
HIV-1 genotyping was performed at the Kenya Medical Research Institute (KEMRI)/CDC HIV laboratory, which has been accredited as a National Drug Resistance Laboratory by the WHO HIV/ResNet group.
ViroSeq HIV-1 genotyping system.
The FDA-approved ViroSeq assay amplifies an 1.8-kb fragment covering the entire protease region and codons 1 to 335 of the reverse transcriptase (RT) region. This assay has an amplification detection sensitivity of 2,000 RNA copies/ml of plasma and a DR-associated mutation (DRM) detection sensitivity and specificity of 99.65% and 99.95%, respectively (35). Drug resistance genotyping was performed according to the manufacturer's instructions, and DR interpretations were conducted using both the ViroSeq HIV-1 genotyping system software, V2.6, and the Stanford genotyping resistance interpretation algorithm (http://sierra2.stanford.edu/sierra/servlet/JSierra) to allow for comparison with the in-house assay.
In-house genotyping assay.
The in-house assay amplifies a 1,084-bp fragment of the HIV-1 pol gene encoding amino acids 6 to 99 of the protease region and codons 1 to 251 of the reverse transcriptase (RT) region. This assay was developed by the CDC, Atlanta, GA, for genotyping all HIV-1 group M subtypes and circulating recombinant forms (CRFs) and has been validated using samples collected from several PEPFAR-supported countries with multiple HIV-1 subtypes and CRFs (30, 32, 33). For this validation, we followed the procedure described by Yang et al. (32). Briefly, 15 μl of nucleic acid extracts from either plasma or DBS was subjected to a one-step RT-PCR using PRTM-F1, which is a 1:1 (wt/wt) combination of two forward primers (F1a, 5′-TGAARGAITGYACTGARAGRCAGGCTAAT, nucleotides [nt] 2057 to 2085 based on HXBII, and F1b, 5′-ACTGARAGRCAGGCTAATTTTTTAG, nt 2068 to 2092), and RT-R1 (reverse, 5′-ATCCCTGCATAAATCTGACTTGC, nt 3370 to 3348) (33; C. Yang, Z. Zhou, J. R. DeVos, and N. Wager, U.S. patent application 61/504,522) and the SuperScript III one-step RT-PCR system with Platinum Taq high-fidelity polymerase according to the manufacturer's protocol (Invitrogen, Carlsbad, CA). For the nested PCR, 4 μl of the RT-PCR product was then used with the inner primers PRT-F2 (forward, 5′-CTTTARCTTCCCTCARATCACTCT, nt 2243 to 2266) and RT-R2 (reverse, 5′-CTTCTGTATGTCATTGACAGTCC, nt 3326 to 3304) (32, 33) to yield an approximately 1.1-kb amplicon. Sequencing was then performed using six overlapping primers, and sequences were analyzed using the ABI 3100 genetic analyzer. Confirmation of base-calling and sequence editing were conducted using the Sequencher V4.5 (Genecodes) software. DR interpretation was performed using the Stanford genotyping drug resistance interpretation algorithm (v4.2.6) (http://sierra2.stanford.edu/sierra/servlet/JSierra) and using the International AIDS Society (IAS) 2011 mutation list (36) for confirmation.
Assay validation.
Performance characteristics of this in-house assay were determined using the WHO/HIV ResNet guidelines for validation of in-house genotyping assays, which circumvent the lack of standard or reference methods for evaluating genotyping performance of DR assays (31). This included assessment of accuracy, precision, reproducibility, and amplification sensitivity as minimal requirements; linearity and sensitivity, though not assessed in this validation, are also considered.
Accuracy.
To assess accuracy, 53 plasma and 52 DBS sample sequences obtained from the in-house assay were compared to the plasma sample sequences obtained from ViroSeq. Accuracy was determined by the degree of concordance between the 66 DRMs identified by ViroSeq and the in-house assay based on the IAS 2011 mutation list (36).
Sensitivity.
Sensitivity was assessed using 68 samples with VL ranges from 253 to 3,264,850 copies/ml, and the genotyping sensitivity is defined as the VL copy ranges at which ≥95% of the samples were successfully genotyped.
Precision and reproducibility.
Precision was assessed using 10 samples with 3 to 5 replicates (5 samples with 5 replicates and 5 samples with 3 replicates), and the reproducibility was determined using 12 samples with at least 2 replicates (5 samples with 5 replicates and 7 duplicate samples). In addition, seven DBS replicates from a 14-member proficiency testing panel from VQA at Chicago, IL, were used to assess the reproducibility for DBS specimens. The replicates in the panel had been shipped under two different temperature conditions. The reproducibility test was performed by two technicians and in some cases utilizing different kit lots. Both precision and reproducibility were determined by the degree of concordance of DR-associated mutations as well as the degree of nucleotide sequence identity.
Contribution of proviral DNA from DBS.
The frequency of amplification of proviral DNA from DBS was investigated using the in-house assay on eight samples with VLs ranging from 15,508 to 3,264,850 copies/ml. These 8 samples were run in the presence and absence of the RT-PCR murine leukemia virus (MuLV) enzyme.
Phylogenetic analysis.
Phylogenetic analysis was performed using the neighbor-joining method in MEGA4 software (37). The evolutionary distances were computed using the maximum composite likelihood method in units of the number of base substitutions per site. The tree was then generated by the neighbor-joining method from a nucleotide alignment of 751 positions devoid of gaps, and tree topology was confirmed by bootstrapping analysis using 1,000 replicates. Pairwise alignment to assess nucleotide sequence identity between matched plasma and DBS and those plasma sequences obtained from the ViroSeq assay was performed using the EMBOSS pairwise alignment tool (http://www.ebi.ac.uk/Tools/psa/emboss_needle/) in the needle global method for whole-length alignment (with default gap-penalty values).
Statistical analysis.
The sensitivity and specificity of the in-house assay for both plasma and DBS were assessed against the ViroSeq system on 66 DRMs as identified in the IAS 2011 mutation list (36). Quantitative variables are expressed as means ± standard deviations (SDs) unless otherwise stated. The McNemar test was then used to assess for significance in the discordant mutations between the in-house and the ViroSeq assays for both sample types. Precision and reproducibility were assessed using the Cohen kappa statistic. Agreement was interpreted as weak (0.400 > κ ≥ 0.200), moderate (0.600 > κ ≥ 0.400), strong (0.800 > κ ≥ 0.600), nearly perfect (1.00 > κ ≥ 0.800), and perfect (κ = 1.000). Sample size was estimated using Buderer's formula (38) with the following assumptions: HIVDR prevalence of 9.3% based on a previous study in Kenya (39), anticipated sensitivity equivalent to that obtained with the ViroSeq system of 99.65 (35), and a two-sided test with an alpha value of 0.05 and a precision value of 0.05. A minimum sample size of 60 was then required in order to obtain a sensitivity equivalent to that of the ViroSeq system.
Quality control and assurance.
The quality of the test runs was ensured by the inclusion of positive and negative controls in each run. Sequence quality was confirmed by using the WHO sequence quality assessment tool (SQUAT) (40) to screen ambiguous nucleotides, frameshifts, insertions, and deletions of the sequences. In addition, ambiguous and atypical amino acids, stop codons, and frameshifts were also screened at the amino acid level. SQUAT was also used to check for cross-contamination by calculating pairwise genetic distances which were further cross-checked by the PAUP tool (41). The KEMRI/CDC HIV research laboratory, where the analysis was performed, participates twice a year in external quality assurance (EQA) programs offered by VQA and the College of American Pathologists (CAP) for HIV plasma genotyping for both the in-house and ViroSeq assays. At the time that this study was conducted, the performances of all the proficiency testing panels were satisfactory. The laboratory is also accredited with the ISO 15189 standard and by WHO ResNet for HIVDR genotyping.
Ethical considerations.
Ethical review committees of the Kenya Medical Research Institute (KEMRI) and the Institutional Review Board of the U.S. Centers for Disease Control and Prevention (CDC), Atlanta, GA, approved this study. All mothers in the KiBS study provided written informed consent that included parental consent for their infants.
Nucleotide sequence accession numbers.
Sequences from this study were submitted to GenBank, and their accession numbers are JQ914045 to JQ914103.
RESULTS
Viral load determination.
Among the 68 plasma samples, the mean plasma VL was 330,855 copies/ml with a median of 58,187 copies/ml, ranging from 253 to 3,264,850 copies/ml. Of these, 44 had VLs of >10,000 copies/ml (median, 374,074), 13 ranged from 1,000 to 10,000 copies/ml (median, 4,040), and the remaining 11 had <1,000 copies/ml (median, 632).
Plasma genotyping sensitivity.
Of the 68 plasma samples used to assess genotyping sensitivity, 100% (95% confidence interval [CI], 93.5 to 100) of all the 44 samples with VLs of >10,000 were genotyped by both assays. All the 13 samples with VLs between 1,000 and 10,000 copies/ml were genotyped with the in-house assay (100%; 95% CI, 80.7 to 100), and 8 (61.5%; 95% CI, 36.1 to 83.1) were genotyped with the ViroSeq assay. For those 11 samples with VLs of <1,000 copies/ml, the in-house assay was able to genotype 7 (63.6%; 95% CI, 36.2 to 85.9) while 1 (9.1%) was genotyped by the ViroSeq system (Table 1).
Table 1.
Patient characteristics and genotyping efficiency from plasma and DBS using the in-house assay compared with the ViroSeq commercial assay
| IDa | Plasma VL (copies/ml) | ART (regimen)b | Length of DBS storage (mo) | Plasma genotype |
DBS genotype (in-house) | HIV-1 subtype | |
|---|---|---|---|---|---|---|---|
| ViroSeq | In-house | ||||||
| 1-0472-8v7 | 3,264,850 | AZT + ABC + KAL | 0–12 | + | + | + | A |
| 1-0079-v8 | 1,408,848 | 3TC + AZT + NVP | 37–48 | + | + | + | A |
| 1-0119-4v11 | 1,220,862 | TN | 25–36 | + | + | + | AD |
| 1-0230-2v7 | 1,133,811 | 3TC + AZT + NVP | 25–36 | + | + | + | CRF10_CD |
| 1-0119-4v7 | 1,023,877 | TN | 25–36 | + | + | + | AD |
| 1-0357-6v11 | 961,300 | AZT + ABC + KAL | 0–12 | + | + | + | A |
| 1-0457-9v5 | 834,528 | TN | 13–24 | + | + | + | A |
| 1-0410-4v9 | 821,222 | AZT + ABC + KAL | 0–12 | + | + | + | AD |
| 1-0437-5v7 | 810,648 | TN | 0–12 | + | + | + | A |
| 1-0357-6v8 | 718,896 | 3TC + AZT + NVP | 13–24 | + | + | + | A |
| 0-0140-4v17 | 714,157 | 3TC + AZT + NVP | 25–36 | + | + | + | A |
| 1-0085-1v8 | 708,158 | 3TC + AZT + NVP | 37–48 | + | + | + | C |
| 1-0496-6v7 | 619,222 | TN | 0–12 | + | + | + | A |
| 1-0360-1v7 | 587,486 | 3TC + AZT + NFV | 13–24 | + | + | + | A |
| 1-0053-3v11 | 562,910 | 3TC + AZT + NVP | 25–36 | + | + | + | A |
| 1-0105-8v12 | 551,800 | 3TC + AZT + NVP | 13–24 | + | + | + | A |
| 1-0437-5v11 | 530,000 | TN | 0–12 | + | + | + | A |
| 1-0066-8v9 | 467,363 | 3TC + AZT + NVP | 25–36 | + | + | + | A |
| 1-0144-5v12 | 431,147 | 3TC + AZT + NVP | 13–24 | + | + | + | D |
| 0-0106-9v16 | 418,017 | 3TC + AZT + NVP | 25–36 | + | + | + | A |
| 0-0137-6v16 | 402,754 | 3TC + AZT + NVP | 37–48 | + | + | + | CRF10_CD |
| 0-0056-6v16 | 380,117 | 3TC + AZT + NVP | 37–48 | + | + | + | C |
| 1-0230-2v5 | 368,032 | TN | 37–48 | + | + | + | CRF10_CD |
| 0-0074-8v16 | 323,445 | 3TC + AZT + NVP | 37–48 | + | + | + | A |
| 1-0410-4v7 | 320,945 | TN | 13–24 | + | + | + | AD |
| 1-0496-6v3 | 303,144 | TN | 13–24 | + | + | + | A |
| 0-0158-1v16 | 292,850 | 3TC + AZT + NVP | 37–48 | + | + | + | A |
| 0-0113-8v17 | 290,549 | 3TC + AZT + NVP | 25–36 | + | + | + | A |
| 1-0317-8v8 | 192,450 | TN | 13–24 | + | + | + | CRF10_CD |
| 0-0150-3v16 | 117,493 | 3TC + AZT + NVP | 37–48 | + | + | + | A |
| 0-0230-2v20 | 111,021 | 3TC + AZT + NVP | 13–24 | + | + | + | CRF10_CD |
| 0-0472-8v3 | 106,234 | 3TC + AZT + NFV | 13–24 | + | + | + | A |
| 1-0289-1v5 | 87,142 | 25–36 | + | + | + | A | |
| 0-0200-6v17 | 72,112 | 3TC + AZT + NVP | 25–36 | + | + | + | A |
| 0-0266-4v16 | 44,262 | 3TC + AZT + NFV | 25–36 | + | + | + | A |
| 0-0496-6v16 | 31,511 | 3TC + AZT + NFV | 0–12 | + | + | + | A |
| 0-0113-8v16 | 30,889 | 3TC + AZT + NVP | 37–48 | + | + | + | A |
| 1-0066-8v1 | 26,412 | TN | 37–48 | + | + | + | A |
| 0-0200-6v16 | 25,395 | 3TC + AZT + NVP | 25–36 | + | + | + | A |
| 0-0127-4v16 | 21,431 | 3TC + AZT + NVP | 37–48 | + | + | + | D |
| 1-0317-8v12 | 21,000 | 3TC + AZT + NFV | 0–12 | + | + | + | CRF10_CD |
| 1-0066-8v7 | 17,591 | 3TC + AZT + NVP | 37–48 | + | + | + | A |
| 0-0182-1v16 | 15,508 | 3TC + AZT + NVP | 37–48 | + | + | + | A |
| 0-0472-8v17 | 11,381 | 3TC + AZT + NFV | 0–12 | + | + | + | A |
| 1-0079-3v5 | 7,398 | TN | 37–48 | + | + | + | A |
| 0-0472-8v16 | 6,923 | 3TC + AZT + NFV | 0–12 | + | + | + | A |
| 0-0157-0v16 | 6,084 | 3TC + AZT + NVP | 37–48 | + | + | + | A |
| 0-0073-7v16 | 5,547 | 3TC + AZT + NVP | 37–48 | + | + | + | A |
| 0-0120-7v16 | 5,434 | 3TC + AZT + NVP | 37–48 | + | + | − | D |
| 1-0517-4v5 | 5,148 | TN | 13–24 | + | + | + | A |
| 1-0360-1v12 | 4,040 | AZT + ABC + KAL | 0–12 | + | + | + | A |
| 0-0036-2v16 | 1,365 | 3TC + AZT + NVP | 37–48 | − | + | − | A |
| 0-0264-2v16 | 1,313 | 3TC + AZT + NVP | 25–36 | − | + | − | A |
| 0-0020-4v16 | 1,106 | 3TC + AZT + NVP | 37–48 | + | + | + | A |
| 0-0517-4v16 | 1,026 | 3TC + AZT + NFV | 0–12 | − | + | − | A |
| 0-0519-6v16 | 1,024 | 3TC + AZT + NFV | 0–12 | − | + | − | G |
| 0-0100-3v16 | 1,006 | 3TC + AZT + NVP | 37–48 | − | + | − | A |
| 0-0475-1v16 | 961 | 3TC + AZT + NFV | 0–12 | − | + | − | A |
| 0-0433-1v16 | 891 | 3TC + AZT + NFV | 13–24 | + | + | + | A |
| 0-0212-0v16 | 792 | 3TC + AZT + NVP | 37–48 | − | + | − | A |
| 0-0093-1v16 | 737 | 3TC + AZT + NVP | 37–48 | − | + | − | A |
| 0-0042-0v16 | 666 | 3TC + AZT + NVP | 37–48 | − | + | − | A |
| 0-0523-2v17 | 632 | 3TC + AZT + NVP | 0–12 | − | + | − | D |
| 0-0227-7v16 | 475 | 3TC + AZT + NVP | 37–48 | − | + | − | A |
| 0-0291-5v16 | 373 | 3TC + AZT + NFV | 13–24 | − | − | − | |
| 0-0276-6v16 | 352 | 3TC + AZT + NFV | 13–24 | − | − | − | |
| 0-0024-8v16 | 260 | 3TC + AZT + NVP | 37–48 | − | − | − | |
| 0-0500-5v16 | 253 | 3TC + AZT + NFV | 0–12 | − | − | − | |
Study sample identifiers (IDs) with specific visit codes (shown by “v”). Patient identifiers starting with 0 are samples collected from the mothers, while those starting with 1 are samples collected from infants. Some of the samples included in the study were collected from the same participants at different study visit points.
Abbreviations: TN, treatment naïve; 3TC, lamivudine; AZT, zidovudine; NVP, nevirapine; KAL, lopinavir-ritonavir (Kaletra); ABC, abacavir; NFV, nelfinavir.
DBS genotyping sensitivity.
After establishing better sensitivity for plasma samples with lower VL measurements in the in-house assay than in the ViroSeq assay, the in-house assay was then used to assess genotyping sensitivity using the matched 44 DBS samples. All the 44 DBS samples (100%; 95% CI, 93.5 to 100) with VLs of >10,000 copies/ml were genotyped by the in-house assay. In addition, 7 (53.8%; 95% CI, 29.1 to 77.3) of the 13 DBS samples with VLs between 1,000 and 10,000 copies/ml were also genotyped, while only 1 of the 11 DBS samples with VLs of <1,000 copies/ml was genotyped (Table 1).
Concordance of the two assays in detecting drug resistance-associated mutations in plasma samples.
The accuracy in detecting DRMs in plasma was determined by using the 53 sequences generated by the ViroSeq system. Two hundred thirty-four DRMs were observed in these 53 sequences, including 68 DRMs in the RT gene and one major and 165 minor mutations in the protease gene. Of these DRMs, 230 were detected in the sequences generated by the in-house assay, which yields an analytic accuracy of 98.29% (95% CI, 97.86 to 98.72) (Tables 2 and 3). Of the four discordant DRMs between the two assays, three were due to mixtures in the RT gene. M184IV and M184MV were detected as mixtures in the ViroSeq system but were nonmixture mutations in the in-house assay (M184V). In contrast, G190AG was detected as a mixture in the in-house assay, but it was a nonmixture in the ViroSeq system (G190A). The one remaining discordant mutation was a minor mutation (G16E) in the protease gene, which was missed by the in-house assay. None of these discordant DRMs were of clinical significance when the sequences were analyzed with the HIValg program using the HIVdb v6.2.0 program deployed at the Stanford HIV Drug Resistance Database. Pairwise comparison of nucleotide substitutions at the DRM sites also indicated high concordance, with only five (0.7%) nonsynonymous substitutions. Overall, compared to the ViroSeq system, the in-house assay gave high specificity in detection of DRMs (99.97%; 95% CI, 99.91 to 100.03) as well as excellent positive and negative predictive values (99.57%; 95% CI, 99.35 to 99.78; 99.88%; 95% CI, 99.76 to 99.99, respectively) (Table 3). Statistical analyses using the McNemar test on paired results for DRMs also supported the excellent performance of the in-house assay compared to the ViroSeq system (χ2 = 1.80, P = 0.375). The excellent performance by the in-house assay was further supported by stratification analysis of DRMs for sensitivity and specificity (sensitivity χ2 = 4, P = 0.125, and specificity χ2 = 1.0, P = 1.0). At the nucleotide level, the mean nucleotide sequence identity was 99.5% for the in-house assay compared to the ViroSeq system. The minor differences observed in sequence identity were mainly caused by mixture bases. We identified 126 mixture base differences among the 53 sequences analyzed; of these, 93 (74%) were detected by the in-house assay and 86 (70%) were detected by the ViroSeq system.
Table 2.
Comparison of drug resistance mutations identified by the in-house assay from plasma and DBS specimens with those identified by ViroSeq from plasma specimens
| IDb | Mutation(s)a |
|||||
|---|---|---|---|---|---|---|
| Protease gene |
Reverse transcriptase gene |
|||||
| ViroSeq plasma | In-house plasma | In-house DBS | ViroSeq plasma | In-house plasma | In-house DBS | |
| 0-0020-4v16 | K20R, M36I, H69K | K20R, M36I, H69K | K20R, M36I, H69K | |||
| 0-0056-6v16 | M36I, H69K | M36I, H69K | M36I, H69K | K103N | K103N | K103N |
| 0-0073-7v16 | G16E, M36I, H69K | G16E, M36I, H69K | G16E, M36I, H69K | V108I, Y181L, M184V | V108I, Y181L, M184V | V108I, Y181L, M184V |
| 0-0074-8v16 | K20R, M36I, H69K | K20R, M36I, H69K | K20R, M36I, H69K | G190A | G190A | G190A |
| 0-0106-9v16 | M36I, H69K | M36I, H69K | M36I, H69K | M184V | M184V | M184V |
| 0-0113-8v16 | M36I, L63P, H69K | M36I, L63P, H69K | M36I, L63P, H69K | K103N, M184V | K103N, M184V | K103N, M184V |
| 0-0113-8v17 | M36I, L63P, H69K | M36I, L63P, H69K | M36I, L63P, H69K | |||
| 0-0120-7v16 | M36I, L63P, V77I | M36I, L63P, V77I | ||||
| 0-0127-4v16 | K20R, M36I, L63P, H69K | K20R, M36I, L63P, H69K | K20R, M36I, L63P, H69K | |||
| 0-0137-6v16 | D60E, I64V, V77I | D60E, I64V, V77I | D60E, I64V, V77I | |||
| 0-0140-4v17 | M36I, M46L, L63LP, H69K | M36I, M46L, L63LP, H69K | M36I, M46L, L63LP, H69K | |||
| 0-0150-3v16 | M36I, H69K | M36I, H69K | M36I, H69K | |||
| 0-0157-0v16 | M36I, K20R, H69K | M36I, K20R, H69K | M36I, K20R, H69K | |||
| 0-0158-1v16 | M36I, H69K | M36I, H69K | M36I, H69K | K103KN | K103KN | K103KN |
| 0-0182-1v16 | K20R, M36I, H69K | K20R, M36I, H69K | K20R, M36I, H69K | M184V, Y188L | M184V, Y188L | M184V, Y188L |
| 0-0200-6v16 | K20R, M36I, H69K | K20R, M36I, H69K | K20R, M36I, H69K | Y181C, M184V | Y181C, M184V | Y181CY, M184IMV |
| 0-0200-6v17 | K20R, M36I, H69K | K20R, M36I, H69K | K20R, M36I, H69K | |||
| 0-0230-2v20 | K20R, M36I, L63P, I64V | K20R, M36I, L63P, I64V | K20R, M36I, L63P, I64V | K103N | K103N | K103N |
| 0-0266-4v16 | G16E, M36I, H69K | G16E, M36I, H69K | G16E, M36I, H69K | |||
| 0-0433-1v16 | M36I, H69K | M36I, H69K | M36I, H69K | M184V | M184V | M184V |
| 0-0472-8v3 | G16E, K20KR, M36I, H69K | G16E, K20KR, M36I, H69K | G16E, K20R, M36I, H69K | |||
| 0-0472-8v16 | G16E, K20R, M36I, H69K | G16E, K20R, M36I, H69K | G16E, K20R, M36I, H69K | K103N, M184MV | K103N, M184V | K103N, Y181CY, M184MV |
| 0-0472-8v17 | G16E, K20R, M36I, H69K | G16E, K20R, M36I, H69K | G16E, K20R, M36I, H69K | K103N, M184V | K103N, M184V | K103KN, M184MV |
| 0-0496-6v16 | M36I, H69K | M36I, H69K | M36I, H69K | |||
| 1-0053-3v11 | M36I, H69K | M36I, H69K | M36I, H69K | |||
| 1-0066-8v1 | K20R, M36I, D60E, I62V, H69K | K20R, M36I, D60E, I62V, H69K | K20R, M36I, D60E, I62V, H69K | |||
| 1-0066-8v7 | K20R, M36I, D60E, I62V, H69K | K20R, M36I, D60E, I62V, H69K | K20R, M36I, D60E, I62V, H69K | D67DN, G190A, T215F | D67DN, G190A, T215F | D67DN, G190A, T215F |
| 1-0066-8v9 | K20R, M36I, D60E, I62V, H69K | K20R, M36I, D60E, I62V, H69K | K20R, M36I, D60E, I62V, H69K | D67DN, K70KR, K101KQ, K103KN, M184V, G190AG, T215FS, K219EK | D67DN, K70KR, K101KQ, K103KN, M184V, G190AG, T215FS, K219EK | D67DN, K70KR, K101KQ, K103KN, M184V, G190AG, T215FIST, K219EK |
| 1-0079-3v5 | G16E, M36I, H69K | G16E, M36I, H69K | G16E, M36I, H69K | K65KR, K101EK, Y181CY, M184MV, G190A | K65KR, K101EK, Y181CY, M184MV, G190A | K65KR, K101EK, Y181CY, M184IMV, G190A |
| 1-0079-v8 | G16E, M36I, H69K | G16E, M36I, H69K | G16E, M36I, H69K | K101E, M184V, G190A | K101E, M184V, G190A | K101E, M184V, G190A |
| 1-0085-1v8 | M36I, H69K, I93L | M36I, H69K, I93L | M36I, H69K, I93L | |||
| 1-0105-8v12 | M36I, D60E, I62V, H69K | M36I, D60E, I62V, H69K | M36I, D60E, I62V, H69K | |||
| 1-0119-4v7 | M36I, H69K | M36I, H69K | M36I, H69K | |||
| 1-0119-4v11 | M36I, H69K | M36I, H69K | M36I, H69K | |||
| 1-0144-5v12 | M36I, I64V | M36I, I64V | M36I, I64V | |||
| 1-0230-2v7 | K20R, M36I, R41K, L63P, I64V | K20R, M36I, R41K, L63P, I64V | K20R, M36I, R41K, L63P, I64V | M184V | M184V | M184V |
| 1-0230-2v7 | K20R, M36I, L63P, I64VF | K20R, M36I, L63P, I64V | K20R, M36I, L63P, I64V | M184V | M184V | M184V |
| 1-0289-1v5 | M36I, H69K | M36I, H69K | M36I, H69K | M184V | M184V | M184V |
| 1-0317-8v8 | M36I, I62V, I64V | M36I, I62V, I64V | M36I, I62V, I64V | Y181C, M184V | Y181C, M184V | Y181C, M184V |
| 1-0317-8v12 | M36I, I62V, I64V | M36I, I62V, I64V | M36IV, I62V, I64V | K70KR, Y181C, M184V | K70KR, Y181C, M184V | Y181C, M184V |
| 1-0357-6v8 | M36I, H69K | M36I, H69K | M36I, H69K | M184V, G190A | M184V, G190AG | M184V, G190A |
| 1-0357-6v11 | M36I, D60E, H69K | M36I, D60E, H69K | M36I, D60E, H69K | A98AG, K103S, M184V, G190A, T215Y | A98AG, K103S, M184V, G190A, T215Y | A98AG, K103S, M184V, G190A, T215Y |
| 1-0360-1v7 | G16E, K20R, M36I, I62V, H69K | G16E, K20R, M36I, I62V, H69K | G16E, K20R, M36I, I62V, H69K | M184V | M184V | M184V |
| 1-0360-1v12 | G16E, K20R, M36I, I62V, H69K | G16E, K20R, M36I, I62V, H69K | G16E, K20R, M36I, I62V, H69K | M184MV | ||
| 1-0410-4v7 | M36I, L63P, H69K | M36I, L63P, H69K | M36I, L63P, H69K | M184MV, T215Y | M184MV, T215Y | M184MV, T215Y |
| 1-0410-4v9 | M36I, L63P, H69K | M36I, L63P, H69K | M36I, L63P, H69K | T215D | T215D | T215D |
| 1-0437-5v7 | M36I, L63P, H69K | M36I, L63P, H69K | M36I, L63P, H69K | M184IV | M184V | M184V |
| 1-0437-5v11 | M36I, L63P, H69K | M36I, L63P, H69K | M36I, L63P, H69K | |||
| 1-0457-9v5 | K20R, M36I, H69K | K20R, M36I, H69K | K20R, M36I, H69K | K65KR, M184MV | K65KR, M184MV | K65KR, M184MV |
| 1-0472-8v7 | G16E, K20R, H69K | G16E, K20R, H69K | G16E, K20R, H69K | K65KR, K103N, Y181CY, M184MV | K65KR, K103N, Y181CY, M184MV | K65KR, K103N, Y181CY, M184V |
| 1-0496-6v3 | G16E, K20R, M36I, H69K | K20R, M36I, H69K | K20R, M36I, H69K | K65KR, M184MV | K65KR, M184MV | K65KR, M184MV |
| 1-0496-6v7 | M36I, H69K | M36I, H69K | M36I, H69K | M184V | M184V | M184V |
| 1-0517-4v5 | M36I, H69K | M36I, H69K | M36I, H69K | M184V | M184V | M184V |
Major drug resistance mutations against protease and reverse transcriptase inhibitors are shown in bold, and discordant drug resistance mutations are underlined.
ID, identifier.
Table 3.
Performance characteristics of the in-house genotyping assay compared to the ViroSeq system in detection of drug resistance mutations
| In-house assay (n) | No. of DRMs by ViroSeq result |
% (95% CI)a |
Mean nucleotide score (%) | ||||
|---|---|---|---|---|---|---|---|
| Positive | Negative | Sensitivity | Specificity | PPV | NPV | ||
| Plasma (53) | 98.29 (97.86–98.72) | 99.97 (99.91–100.03) | 99.57 (99.35–99.78) | 99.88 (99.76–99.99) | 99.5 | ||
| Positive | 230 | 1 | |||||
| Negative | 1 | 3,263 | |||||
| DBS (52) | 96.54 (95.93–97.15) | 99.97 (99.91–100.03) | 99.55 (99.33–99.78) | 99.75 (99.58–99.92) | 99.5 | ||
| Positive | 223 | 1 | |||||
| Negative | 8 | 3,200 | |||||
CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value.
Concordance of the two assays in detecting drug resistance mutations in DBS samples.
Based on the satisfactory results obtained with the in-house assay in comparison with the ViroSeq system as well as the higher sensitivity in genotyping plasma samples with lower VLs with the in-house assay, we next genotyped and analyzed the 52 matched DBS samples using the in-house assay only and compared the genotyping results with those obtained from plasma samples by the ViroSeq system. Compared to sequences obtained from plasma samples using the ViroSeq system, the in-house assay gave a sensitivity of 96.54% (95% CI, 95.93 to 97.15) in DRM detection (Tables 2 and 3). A total of 231 mutations were observed in the ViroSeq system. Of these, 223 were identified in DBS sequences by the in-house assay. Among the eight discordant DRMs, five occurred in the RT gene, four of which were caused by mixtures (Y181CY, K103KN, T215FIST, and MI84IMV) while the remaining one (K70KR) was not detected in the DBS sample by the in-house assay. The remaining three discordant mutations were detected in the protease gene, two of which occurred as mixtures at minor DR positions (K20KR and M36IV) while the other one was absent (G16E) in the in-house assay. In addition, one extra DRM (Y181CY) was present only in one DBS sample. Despite the DRM detection differences, they were not clinically significant even for the patient with the Y181CY mutation since this patient had already had a K013KN mutation which led to high-level resistance to nonnucleoside reverse transcriptase inhibitors (NNRTI). Pairwise comparison of nucleotide substitution at DRM sites was also highly concordant for DBS with only 12 (1.7%) nonsynonymous substitutions. Similar to plasma, the in-house assay also gave near-perfect specificity in detecting DRMs in DBS samples (99.97%; 95% CI, 99.91 to 100.03) with excellent positive (99.55; 95% CI, 99.33 to 99.78) and negative (99.75; 95% CI, 99.58 to 99.92) predictive values (Table 3). McNemar test analysis on paired results for DRM detection confirmed the excellent concordant results (χ2 = 5.44, P = 0.04) and great sensitivity and specificity (χ2 = 8, P = 0.008, and χ2 = 1.0, P = 1.0), respectively. As expected, at the nucleotide level, the in-house assay also gave excellent mean nucleotide identity compared to the plasma ViroSeq system (99.5%). The minor differences from the McNemar test as well as nucleotide sequence identity were caused by mixture bases, of which the in-house assay using DBS was able to detect 79 of 126 (63%), compared to 70% for the ViroSeq system.
Precision and reproducibility.
Using the WHO criteria for DBS genotyping method validation, the in-house assay demonstrated a high precision and reproducibility in both DBS and plasma specimen types (Table 4). Overall agreement for precision was nearly perfect with a kappa score of 0.974 (95% CI, 0.960 to 0.989) for plasma and 0.967 (95% CI, 0.960 to 0.968) for DBS. The mean nucleotide sequence identity score for precision was 99.4% ± 0.33% in plasma and 99.2% ± 0.59% in DBS. Similar results were also obtained in the reproducibility assessment, where the kappa score was nearly perfect: 0.975 (95% CI, 0.956 to 0.980) for plasma and 0.992 (95% CI, 0.991 to 0.995) for DBS. The mean nucleotide sequence identity was 99.3% ± 0.35% for plasma and 99.1% ± 0.59% for DBS.
Table 4.
Precision and reproducibility of genotyping using plasma and DBS with the in-house assay
| Quality | No. of samples | No. of replicates | κ value | 95% CI | Interpretationa | Mean nucleotide identity score ± SD (%) |
|---|---|---|---|---|---|---|
| Precision | ||||||
| Plasma | 10 | 40 | 0.974 | 0.960–0.989 | NP | 99.4 ± 0.33 |
| DBS | 10 | 40 | 0.967 | 0.960–0.968 | NP | 99.2 ± 0.59 |
| Reproducibility | ||||||
| Plasma | 10 | 39 | 0.975 | 0.956–0.980 | NP | 99.3 ± 0.35 |
| DBS | 10 | 53 | 0.992 | 0.991–0.995 | NP | 99.1 ± 0.59 |
Interpretation for kappa statistic; NP, near perfect.
Frequency of proviral DNA amplification from DBS.
We also assessed the amplification contribution of proviral DNA in the TNA obtained from DBS using 8 samples with VLs ranging from 15,508 to 3,264,850 copies/ml. These 8 samples were run in the presence and absence of the RT-PCR MuLV enzyme. In the presence of the reverse transcription enzyme, the pol gene was successfully amplified in all eight samples, but only one sample with a higher VL (3,264,850 copies) could be amplified in the absence of the RT-PCR enzyme.
HIV-1 subtype analysis.
Of the 64 newly obtained sequences, 49 were obtained from independent participants and the remaining 15 were obtained at different time points from these participants. Of the 49 samples, 36 were subtype A1 (73.5%), 4 were D (8.2%), 2 were C (4.1%), and 1 was G (2.0%) while 6 (12.2%) were recombinants: 4 CRF10_CD and 2 AD recombinants (Fig. 1).
Fig 1.
Phylogenetic analysis showing correlation of plasma genotypes from the in-house assay and ViroSeq assay and DBS from the in-house assay. IH, in-house; PL, plasma; VS, ViroSeq. The number at the node denotes bootstrap values of greater than 70%.
Comparative analysis of cost and total hands-on time between the two assays using plasma and DBS samples.
Table 5 describes the hands-on time and cost of the different steps involved in the HIVDR testing process, from sample collection to generating results. The cost included reagents and disposables based on 2011 U.S. dollars plus the cost for major equipment maintenance. Fixed costs, such as purchasing equipment and software, personnel, and other logistics, such as those of storage, transportation, and external quality assurance (EQA), were omitted. The cost of running the in-house assay using DBS specimens was $110. 05, that for the in-house assay using plasma specimens was $113.33, and that for the ViroSeq system using plasma specimens was $278.31. In addition, the in-house assay using plasma required the least hands-on time, ∼16 h 30 min, compared to ∼17 h 10 min for ViroSeq and ∼24 h 20 min for the in-house assay using DBS, which required the longest hands-on time due to the manual extraction procedure.
Table 5.
Genotyping cost analysis and estimated hands-on timea
| Process | Step(s) | Assay and sample type |
|||||
|---|---|---|---|---|---|---|---|
| ViroSeq plasma |
In-house plasma |
In-house DBS |
|||||
| Hands-on time | Cost/test ($) | Hands-on time | Cost/test ($) | Hands-on time | Cost/test ($) | ||
| Sample preparation | Sample collection (either DBS or blood in EDTA) and plasma separation | 6 h | 6.00 | 6 h | 6.00 | 10 min | 1.00 |
| Nucleic acid extraction | RNA extraction, plasma | 3 h | 150 | 1 h | 5.13 | ||
| TNA extraction, DBS | 3 h | 6.85 | |||||
| Amplification | RT-PCR | 4 h | 13.96 | 4 h | 13.96 | ||
| One-step RT-PCR | 5 h 20 min | ||||||
| Nested PCR | 3 h | 5.12 | 3 h | 5.12 | |||
| Gel documentation | Gel electrophoresis | 1 h | 1 h | 19.22 | 1 h | 19.22 | |
| Genotyping | Amplicon purification | 50 min | 30 min | 3.59 | 30 min | 3.59 | |
| Sequencing | 2 h 30 min | 90 | 2 h 30 min | 28.0 | 2 h 30 min | 28.0 | |
| Sequence amplicon purification | 1 h 30 min | 13.76 | 1 h 30 min | 13.76 | 1 h 30 min | 13.76 | |
| Sequence detection, visualization | 2 h 30 min | 14.53 | 2 h 30 min | 14.53 | 2 h 30 min | 14.53 | |
| Sequence analysis | Sequence validation, interpretation, and quality analysis | 20 min | 20 min | 20 min | |||
| Equipment maintenance | ABI 3100 and thermocyclers, biosafety cabinets | 4.02 | 4.02 | 4.02 | |||
| Total | 17 h 10 min | 278.31 | 16 h 30 min | 113.33 | 24 h 20 min | 110.05 | |
The cost was estimated based on U.S. dollars at the 2011 market values. The estimated costs also included the costs to run control specimens in each run and equipment maintenance but excluded fixed costs, personnel costs, and other logistics, such as storage, transportation, and enrollment in EQA programs, which are vital for the quality-assured genotyping results.
DISCUSSION
We observed excellent concordance of DRM detections in plasma and DBS by the in-house genotyping assay compared to the ViroSeq assay and confirmed the previous report on the excellent performance of this assay (32, 33). The overall sensitivity and specificity of the in-house assay in detecting DRMs were 98.29% and 99.97% for plasma and 96.54% and 99.97% for DBS, respectively. The findings from this study affirm previous reports that DBS specimens offer an excellent alternative sample type for HIV genotyping for HIVDR tests (13, 29, 32, 42–46). This study further demonstrates their utility and feasibility in a resource-limited setting for use in HIVDR monitoring and surveillance combined with a less costly and subtype-independent in-house genotyping assay. The overall genotyping efficiency rate for the in-house assay was 94% compared to 78% by the ViroSeq system. The relatively higher genotyping rate for the in-house assay on plasma specimens may be attributed to the shorter fragment target (1.1 versus 1.8 kb) as well as the inclusion of a nested PCR step. However, this shorter amplicon has no effect on the interpretation of the known clinically relevant DRMs described in the IAS 2011 mutation list for the protease and RT regions (36).The overall genotyping rate for DBS specimens using the in-house assay was moderate (76%). However, when only those DBS specimens with VLs of ≥1,000 copies/ml were considered according to the WHO recommendation for ART patient monitoring (47), the genotyping rate was greatly improved (89.5%). The lower-than-expected DBS genotyping rate might result in the lower input for TNA extraction in the current study, since we used only two 6-mm discs, which are equivalent to 32 μl of whole blood. A previous study by Masciotra et al. (13) showed that TNA amplification from lower-VL samples can be achieved from DBS by using 2 complete spots (about 100 μl). This was, however, not feasible in this analysis due to the depletion of some of the samples. In fact, the moderate success rate despite lower initial sample input may be a reflection of a higher sensitivity of the in-house assay. Apart from the lower initial input, DBS genotyping efficiency may also be dependent on the storage conditions; in this case, the DBS samples had been stored for a period of up to 4 years under optimal conditions of −30°C. This demonstrates the integrity of the DBS samples in preserving the viral genetic materials for a long period under optimal storage conditions and is consistent with the findings from the work of Masciotra et al. (13). This is vital in HIVDR monitoring surveys in resource-limited settings where testing is usually performed either in batched samples at the centralized genotyping laboratories within the country or in regional laboratories outside the country in some situations.
The sensitivity and specificity of the in-house assay in detecting DRMs were highly concordant with those obtained in the ViroSeq system for both plasma and DBS. The discordant DRMs detected between the in-house and ViroSeq on both DBS and plasma were mainly caused by mixture bases at the DRM sites. These differences, however, were not of clinical significance, as genotypic algorithms infer the presence of resistance for a synonymous mixture mutation in the same way as they do for a complete mutation.
Another important requirement in method validation is the ability to produce accurate results within a run or between runs under similar assay conditions. The in-house assay assessed using the WHO guidelines for genotyping method validation demonstrated both good intra- and interrun comparisons as well as excellent performance in proficiency test panels for both plasma and DBS. This further confirmed a possible concordance of results if this assay is adopted for use in other laboratories in resource-limited settings. In fact, the assay has been implemented in the Ethiopian national DR laboratory and the laboratory has been accredited by the WHO ResNet group as a National DR Laboratory (NDRL) (http://www.who.int/hiv/topics/drugresistance/technical_information/en/index.html).
The high concordance of the DRMs detected by the ViroSeq commercial assay and in-house assay (both plasma and DBS) suggests the utility of this assay in both HIVDR surveillance and monitoring. Twenty-five of the samples used in the analysis were obtained from mothers experiencing virological failures (plasma VLs of ≥1,000 copies/ml) in the KiBS trial of prevention of mother-to-child transmission (PMTCT). The HIVDR status of these mothers had been assessed in real time during the trial using the ViroSeq system; thus, confirmation of the DRM detection using the in-house assay on both plasma and DBS suggests the clinical utility of this assay as an HIVDR monitoring tool using both sample types. However, due to the contribution of proviral DNA from peripheral blood mononuclear cells (PBMCs) in DBS, DBS is not recommended for use in long-term ART-experienced patients, as the detected DRMs may not be a full reflection of the circulating viruses but may reflect archived ones (31). Archived viruses in the DNA have been reported to have different mutation patterns from those of circulating viruses, especially with ART-experienced persons (25, 29, 45). As a result, genotyping from DBS which contain both the proviral DNA and free circulating RNA viruses is expected to give mutation patterns different from those of plasma in ART-experienced persons in the case of the presence of archived mutations. In this study, we observed minimal proviral contributions with only two discordant major mutations: one present in DBS but missing in plasma and vice versa. Of interest was the low frequency of proviral DNA amplification; only 1 of the 8 samples was amplified in the absence of an RT enzyme, with the amplified sample having a very high VL of >3.6 million copies/ml. This may suggest that genotyping from DBS involves mainly the free circulating viral RNA rather than the archived DNA provirus. However, the use of DBS in long-term ART-experienced patients should be made with caution, as the samples in this analysis were obtained from patients who were on ART for <1 year.
The ability to genotype multiple HIV-1 subtypes and CRFs is essential for any genotyping assay in sub-Saharan Africa where multiple subtypes and CRFs exist (32, 48–50). This is specifically important since the two FDA-approved assays were developed and evaluated for use with subtype B viruses (23). Previous studies using the in-house assay have shown its suitability as a broadly sensitive genotyping assay which was able to genotype HIV subtypes C, B, CRF 01 AE and 02 AG, A, and CRF 07 and 08 BC (32, 33). From this analysis, we further affirm its broad subtype sensitivity, especially for subtypes A and D, which are also the predominant HIV strains in Kenya. It is important to note that the ViroSeq system also gave a satisfactory performance in genotyping the different virus strains in this study. However, the lower DBS sensitivity reported before (13) may restrict its utility in resource-limited settings for HIVDR monitoring surveys due to logistical constraints for collecting plasma specimens.
One of the reasons that the HIVDR genotyping test has not become a routine service is the high cost associated with HIVDR monitoring. In this study, we performed analysis on the cost associated with the two assays using the 2011 market values in U.S. dollars. Cost assessments of the in-house assay against the ViroSeq system showed that genotyping of DBS or plasma with the in-house assay could reduce the HIVDR testing cost by about 60%.
This study had some limitations. First, our sample size was relatively small, although it met the calculated sample size for our study purpose. Another limitation of this study is the DBS collection procedure, which was inherited from the original study design, for which the DBS from infants had not been collected by either finger or heel pricks. However, studies using DBS collected from infants by heel/finger pricks are under way, which will add the needed data for the feasibility of using DBS collected by heel/finger pricks for HIVDR testing. Despite the limitations of the study, we believe that there are several strengths. (i) The use of the recently developed WHO guidelines in the validation process not only increases the confidence of the performance of the assay but also serves to ease the adoption by other laboratories in resource-limited settings. (ii) This validation was conducted under real field setup conditions, and it was a true reflection of how this assay could be performed in resource-limited settings. (iii) The implementation of the in-house assay in another resource-limited country and further attainment of WHO ResNet accreditation as mentioned above are another testimony that if vigorous implementation procedures are followed, then the low-cost in-house assay, like the one that we validated here, can be successfully implemented in resource-limited settings. Other considerations like having well-trained, qualified, and competent personnel and meeting the minimal infrastructure requirement as well as participation in external proficiency programs should also be in place.
In conclusion, this study demonstrates the excellent performance of a lower-cost in-house genotyping assay for both plasma and DBS specimens for use in HIVDR surveillance and ART patient monitoring. This assay would be particularly appropriate for use in resource-limited settings with DBS specimens. These findings are of particular interest due to the increased need for HIVDR genotyping in resource-limited settings in the era of increasing demand for ARV usage in treating HIV-infected patients as well as treatment for prevention.
ACKNOWLEDGMENTS
We are grateful to the study participants, the Kisumu Breastfeeding Study, the KEMRI-CDC HIV Research Laboratory, the Kenya Medical Research Institute, and the Kenya Ministry of Health, whose participation made this study possible.
We also thank GlaxoSmithKline and Boehringer Ingelheim for providing the study medications and the Division of HIV/AIDS Prevention, Centers for Disease Control and Prevention, Atlanta, GA, for funding. This research has also been supported by the President's Emergency Plan for AIDS Relief (PEPFAR) through the Centers for Disease Control and Prevention.
This paper is published with the permission of the Director of KEMRI.
The findings and conclusions in this article are those of the authors and do not necessarily represent the views of the U.S. Centers for Disease Control and Prevention. Use of trade names is for identification purposes only and does not constitute endorsement by the U.S. Centers for Disease Control and Prevention or the Department of Health and Human Services.
Footnotes
Published ahead of print 5 December 2012
REFERENCES
- 1. Global Fund To Fight AIDS, Tuberculosis and Malaria 2009. Scaling up for impact: results report. Executive summary. Global Fund To Fight AIDS, Tuberculosis and Malaria, Geneva, Switzerland [Google Scholar]
- 2. PEPFAR 2009. Celebrating life: fifth annual report to Congress on PEPFAR. The United States President's Emergency Plan for AIDS Relief, Washington, DC [Google Scholar]
- 3. UNAIDS 2008. 2008 report on the global AIDS epidemic. UNAIDS, Geneva, Switzerland [Google Scholar]
- 4. WHO, UNAIDS, UNICEF 2009. Towards universal access: scaling up priority HIV/AIDS interventions in the health sector. September 2009 progress report. WHO, Geneva, Switzerland [Google Scholar]
- 5. Bertagnolio S, Derdelinckx I, Parker M, Fitzgibbon J, Fleury H, Peeters M, Schuurman R, Pillay D, Morris L, Tanuri A, Gershy-Damet GM, Nkengasong J, Gilks CF, Sutherland D, Sandstrom P. 2008. World Health Organization/HIVResNet drug resistance laboratory strategy. Antivir. Ther. 13(Suppl. 2):49–57 [PubMed] [Google Scholar]
- 6. Hammer SM, Eron JJ, Jr, Reiss P, Schooley RT, Thompson MA, Walmsley S, Cahn P, Fischl MA, Gatell JM, Hirsch MS, Jacobsen DM, Montaner JS, Richman DD, Yeni PG, Volberding PA. 2008. Antiretroviral treatment of adult HIV infection: 2008 recommendations of the International AIDS Society-USA panel. JAMA 300:555–570 [DOI] [PubMed] [Google Scholar]
- 7. Hirsch MS, Conway B, D'Aquila RT, Johnson VA, Brun-Vezinet F, Clotet B, Demeter LM, Hammer SM, Jacobsen DM, Kuritzkes DR, Loveday C, Mellors JW, Vella S, Richman DD. 1998. Antiretroviral drug resistance testing in adults with HIV infection: implications for clinical management. International AIDS Society-USA panel. JAMA 279:1984–1991 [DOI] [PubMed] [Google Scholar]
- 8. Birx D, de Souza M, Nkengasong JN. 2009. Laboratory challenges in the scaling up of HIV, TB, and malaria programs: the interaction of health and laboratory systems, clinical research, and service delivery. Am. J. Clin. Pathol. 131:849–851 [DOI] [PubMed] [Google Scholar]
- 9. Martinson N, Mohapi L, Bakos D, Gray GE, McIntyre JA, Holmes CB. 2009. Costs of providing care for HIV-infected adults in an urban HIV clinic in Soweto, South Africa. J. Acquir. Immune Defic. Syndr. 50:327–330 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Cecelia AJ, Balakrishnan P, Mohanakrishnan, Venkatakrishnan B, Solomon S, Kumarasamy N. 2006. Effective evaluation of novel low-cost CD4 monitoring assays. J. Immunol. Methods 316:158–162 [DOI] [PubMed] [Google Scholar]
- 11. Fiscus SA, Cheng B, Crowe SM, Demeter L, Jennings C, Miller V, Respess R, Stevens W. 2006. HIV-1 viral load assays for resource-limited settings. PLoS Med. 3:e417 doi:10.1371/journal.pmed.0030417 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Jani IV, Sitoe NE, Chongo PL, Alfai ER, Quevedo JI, Tobaiwa O, Lehe JD, Peter TF. 2011. Accurate CD4 T-cell enumeration and antiretroviral drug toxicity monitoring in primary healthcare clinics using point-of-care testing. AIDS 25:807–812 [DOI] [PubMed] [Google Scholar]
- 13. Masciotra S, Garrido C, Youngpairoj AS, McNulty A, Zahonero N, Corral A, Heneine W, de Mendoza C, Garcia-Lerma JG. 2007. High concordance between HIV-1 drug resistance genotypes generated from plasma and dried blood spots in antiretroviral-experienced patients. AIDS 21:2503–2511 [DOI] [PubMed] [Google Scholar]
- 14. Murtagh MM. 2011. UNITAID technical report: HIV/AIDS diagnostic landscape, 2nd ed UNITAID, Geneva, Switzerland [Google Scholar]
- 15. Mehta N, Trzmielina S, Nonyane BA, Eliot MN, Lin R, Foulkes AS, McNeal K, Ammann A, Eulalievyolo V, Sullivan JL, Luzuriaga K, Somasundaran M. 2009. Low-cost HIV-1 diagnosis and quantification in dried blood spots by real time PCR. PLoS One 4:e5819 doi:10.1371/journal.pone.0005819 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Greengrass V, Plate MM, Steele PM, Taqi R, van Rooijen L, Sun Y, Morris LM, Crowe SM. 2007. HIV monitoring in laboratories in resource limited areas. J. Chin. Clin. Med. 2:469–480 [Google Scholar]
- 17. Bennett DE, Bertagnolio S, Sutherland D, Gilks CF. 2008. The World Health Organization's global strategy for prevention and assessment of HIV drug resistance. Antivir. Ther. 13(Suppl. 2):1–13 [PubMed] [Google Scholar]
- 18. Bergeron M, Daneau G, Ding T, Sitoe NE, Westerman LE, Stokx J, Jani IV, Coetzee LM, Scott L, De Weggheleire A, Boel L, Stevens WS, Glencross DK, Peter TF. 2012. Correction: performance of the PointCare NOW system for CD4 counting in HIV patients based on five independent evaluations. PLoS One 7(9). doi:10.1371/annotation/71a6d947-c034-4749-a33d-971284ea3408 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Bertagnolio S, Sutherland D. 2005. WHO approach to track HIV drug resistance emergence and transmission in countries scaling up HIV treatment. AIDS 19:1329–1330 [DOI] [PubMed] [Google Scholar]
- 20. Bennett DE, Myatt M, Bertagnolio S, Sutherland D, Gilks CF. 2008. Recommendations for surveillance of transmitted HIV drug resistance in countries scaling up antiretroviral treatment. Antivir. Ther. 13(Suppl. 2):25–36 [PubMed] [Google Scholar]
- 21. Cochrane J. 2000. Narrowing the gap: access to HIV treatments in developing countries. A pharmaceutical company's perspective. J. Med. Ethics 26:47–50 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Crowe S, Turnbull S, Oelrichs R, Dunne A. 2003. Monitoring of human immunodeficiency virus infection in resource-constrained countries. Clin. Infect. Dis. 37:S25–S35 [DOI] [PubMed] [Google Scholar]
- 23. Jagodzinski LL, Cooley JD, Weber M, Michael NL. 2003. Performance characteristics of human immunodeficiency virus type 1 (HIV-1) genotyping systems in sequence-based analysis of subtypes other than HIV-1 subtype B. J. Clin. Microbiol. 41:998–1003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Saravanan S, Vidya M, Balakrishnan P, Kumarasamy N, Solomon SS, Solomon S, Kantor R, Katzenstein D, Ramratnam B, Mayer KH. 2009. Evaluation of two human immunodeficiency virus-1 genotyping systems: ViroSeq 2.0 and an in-house method. J. Virol. Methods 159:211–216 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Hallack R, Doherty LE, Wethers JA, Parker MM. 2008. Evaluation of dried blood spot specimens for HIV-1 drug-resistance testing using the Trugene HIV-1 genotyping assay. J. Clin. Virol. 41:283–287 [DOI] [PubMed] [Google Scholar]
- 26. Parkin N, Diallo K, Parry CR, Mwebaza S, Batamwita R, Devos J, Bbosa B, Lyaboga F, Magambo B, Jordan M, Downing R, Kaleebu P, Yang C, Bertagnolio S. 2012. Field study of the utility of dried blood spots (DBS) for HIV-1 drug resistance (HIVDR) genotyping in Kampala, Uganda: storage for 2 weeks and shipping at ambient temperature has no effect on genotyping efficiency. International Workshop on HIV & Hepatitis Virus Drug Resistance and Curative Strategies, vol. 17 International Medical Press, Melia Sitges, Spain [Google Scholar]
- 27. WHO 2010. Guidance on regulations for the transport of infectious substances 2011–2012. WHO/HSE/IHR/2010.8 WHO, Geneva, Switzerland [Google Scholar]
- 28. Buckton AJ, Bissett SL, Myers RE, Beddows S, Edwards S, Cane PA, Pillay D. 2008. Development and optimization of an internally controlled dried blood spot assay for surveillance of human immunodeficiency virus type-1 drug resistance. J. Antimicrob. Chemother. 62:1191–1198 [DOI] [PubMed] [Google Scholar]
- 29. McNulty A, Jennings C, Bennett D, Fitzgibbon J, Bremer JW, Ussery M, Kalish ML, Heneine W, Garcia-Lerma JG. 2007. Evaluation of dried blood spots for human immunodeficiency virus type 1 drug resistance testing. J. Clin. Microbiol. 45:517–521 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Parkin N, de Mendoza C, Schuurman R, Jennings C, Bremer J, Jordan MR, Bertagnolio S. 2012. Evaluation of in-house genotyping assay performance using dried blood spot specimens in the Global World Health Organization laboratory network. Clin. Infect. Dis. 54(Suppl 4):S273–S279 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. WHO 2010. WHO manual for HIV drug resistance testing using dried blood spots specimens. WHO, Geneva, Switzerland [Google Scholar]
- 32. Yang C, McNulty A, Diallo K, Zhang J, Titanji B, Kassim S, Wadonda-Kabondo N, Aberle-Grasse J, Kibuka T, Ndumbe PM, Vedapuri S, Zhou Z, Chilima B, Nkengasong JN. 2010. Development and application of a broadly sensitive dried-blood-spot-based genotyping assay for global surveillance of HIV-1 drug resistance. J. Clin. Microbiol. 48:3158–3164 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Zhou Z, Wagar N, DeVos JR, Rottinghaus E, Diallo K, Nguyen DB, Bassey O, Ugbena R, Wadonda-Kabondo N, McConnell MS, Zulu I, Chilima B, Nkengasong J, Yang C. 2011. Optimization of a low cost and broadly sensitive genotyping assay for HIV-1 drug resistance surveillance and monitoring in resource-limited settings. PLoS One 6:e28184 doi:10.1371/journal.pone.0028184 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Thomas TK, Masaba R, Borkowf CB, Ndivo R, Zeh C, Misore A, Otieno J, Jamieson D, Thigpen MC, Bulterys M, Slutsker L, De Cock KM, Amornkul PN, Greenberg AE, Fowler MG. 2011. Triple-antiretroviral prophylaxis to prevent mother-to-child HIV transmission through breastfeeding—the Kisumu Breastfeeding Study, Kenya: a clinical trial. PLoS Med. 8:e1001015 doi:10.1371/journal.pmed.1001015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Eshleman SH, Crutcher G, Petrauskene O, Kunstman K, Cunningham SP, Trevino C, Davis C, Kennedy J, Fairman J, Foley B, Kop J. 2005. Sensitivity and specificity of the ViroSeq human immunodeficiency virus type 1 (HIV-1) genotyping system for detection of HIV-1 drug resistance mutations by use of an ABI PRISM 3100 genetic analyzer. J. Clin. Microbiol. 43:813–817 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Johnson VA, Calvez V, Gunthard HF, Paredes R, Pillay D, Shafer R, Wensing AM, Richman DD. 2011. 2011 update of the drug resistance mutations in HIV-1. Top. Antivir. Med. 19:156–164 [PMC free article] [PubMed] [Google Scholar]
- 37. Tamura K, Dudley J, Nei M, Kumar S. 2007. MEGA4: molecular evolutionary genetics analysis (MEGA) software version 4.0. Mol. Biol. Evol. 24:1596–1599 [DOI] [PubMed] [Google Scholar]
- 38. Buderer NM. 1996. Statistical methodology: I. Incorporating the prevalence of disease into the sample size calculation for sensitivity and specificity. Acad. Emerg. Med. 3:895–900 [DOI] [PubMed] [Google Scholar]
- 39. Steegen K, Luchters S, Dauwe K, Reynaerts J, Mandaliya K, Jaoko W, Plum J, Temmerman M, Verhofstede C. 2009. Effectiveness of antiretroviral therapy and development of drug resistance in HIV-1 infected patients in Mombasa, Kenya. AIDS Res. Ther. 6:12 doi:10.1186/1742-6405-6-12 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. WHO 2010. Sequence quality assessment tool (SQUAT). WHO, Geneva, Switzerland [Google Scholar]
- 41. Swofford DL. 2003. PAUP*. Phylogenetic analysis using parsimony (*and other methods), version 4. Sinauer Associates, Sunderland, MA [Google Scholar]
- 42. Johannessen A. 2010. Dried blood spots in HIV monitoring: applications in resource-limited settings. Bioanalysis 2:1893–1908 [DOI] [PubMed] [Google Scholar]
- 43. 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] [PubMed] [Google Scholar]
- 44. 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] [PubMed] [Google Scholar]
- 45. Lira R, Valdez-Salazar H, Vazquez-Rosales G, Rojas-Montes O, Ruiz-Tachiquin M, Torres-Ibarra R, Cano-Dominguez C, Maldonado-Rodriguez A, Gomez A, Munoz O, Alvarez-Munoz MT. 2010. Genotypic testing for HIV-1 drug resistance using dried blood samples. Arch. Virol. 155:1117–1125 [DOI] [PubMed] [Google Scholar]
- 46. 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] [PMC free article] [PubMed] [Google Scholar]
- 47. Jordan MR, Bennett DE, Bertagnolio S, Gilks CF, Sutherland D. 2008. World Health Organization surveys to monitor HIV drug resistance prevention and associated factors in sentinel antiretroviral treatment sites. Antivir. Ther. 13(Suppl. 2):15–23 [PubMed] [Google Scholar]
- 48. Kandathil AJ, Ramalingam S, Kannangai R, David S, Sridharan G. 2005. Molecular epidemiology of HIV. Indian J. Med. Res. 121:333–344 [PubMed] [Google Scholar]
- 49. Nkengasong JN, Adje-Toure C, Weidle PJ. 2004. HIV antiretroviral drug resistance in Africa. AIDS Rev. 6:4–12 [PubMed] [Google Scholar]
- 50. Peeters M, Toure-Kane C, Nkengasong JN. 2003. Genetic diversity of HIV in Africa: impact on diagnosis, treatment, vaccine development and trials. AIDS 17:2547–2560 [DOI] [PubMed] [Google Scholar]