Abstract
Background
Successful antiretroviral (ARV) treatment requires both excellent patient compliance and adequate treatment monitoring. Although the access to ARVs is being improved in most of the resource-limited settings most affected by the pandemic, the assessment of treatment success or failure remains a major challenge. Several techniques have been approved for viral load determination, most of which are sophisticated and expensive. One of such techniques, the COBAS Amplicor HIV-1 monitor, was the first automated viral load measurement technique, 1 and can be considered a reference technique. Unfortunately, most resource limited settings lack both the material, human, and financial resources required to implement such a reference standard for monitoring viral load. Less sophisticated and cheaper techniques that are practicable at the regional and peripheral level of health systems in these resource-limited settings are thus imperative. These include, among others, manual methods that do not require sophisticated materials and are probably less expensive. The validity and cost-effectiveness of such manual methods however need to be evaluated before their adoption and widespread use.
In view of assessing the possibility of implementing the manual technique at the regional and peripheral levels of health systems in resource-limited setting, in this study we assess the utility of a manual method of viral load quantification by evaluating its correlation with HIV viral loads determined with an automated method.
Materials and Methods
Patients were recruited from the “Day clinic” of the Yaounde Central Hospital, and the Centre for the Study and Control of Communicable Diseases (CSCCD) of the Faculty of Medicine and Biomedical Sciences in Yaounde, Cameroon. Only HIV sero-positive treatment-naïve patients who gave a formal consent were included in the study. Patients' age, sex, marital status, province of residence, date of HIV diagnosis, test used for HIV diagnosis, CD4 count, and clinical category were recorded on a standardized questionnaire. Patients were bled 5 mL of venous blood, using a Vacutainer needle and an EDTA Vacutainer tube. Plasma was then aliquoted less than 6 hours after bleeding and stored in cryotubes at a temperature of −20°C until analysis. Samples were stored and analyzed at the laboratory of the CSCCD. A total of 40 patients were recruited for the study.
Since our viral load assays could detect only HIV-1 viruses, infection with HIV-1 was initially confirmed using serology: HIV sero-positivity was confirmed from all the samples using the Enzgnost Integral test (Dade, Bahring Germany), an enzyme-linked immunosorbent assay (ELISA) test capable of detecting both antibodies against HIV antigens and also detecting HIV antigens from plasma; after which HIV-1 was diagnosed using the Serodia HIV-1/2 test (Fujirebio Inc., Tokyo, Japan), a serology test that separately identifies infection from HIV-1 and HIV-2.
HIV-1 Viral Load Determination
Manual method using the Amplicor HIV-1 Monitor ultrasensitive method
HIV-1 viral load levels were determined manually using the Amplicor HIV-1 Monitor ultrasensitive method (Roche, Germany). The assay was implemented according to the manufacturer’s manual. Briefly, this is a reverse transcription-polymerase chain reaction (RT-PCR) method of viral load determination. The procedure involves viral RNA extraction from viral particles in plasma. This was done by lysis of viral particles using the products' lysis reagent, and precipitation of the lysate using isopropanol. Reverse transcription and PCR were done in one step in a thermal cycler using a thermostable recombinant enzyme thermus thermophilus DNA polymerase, which has both reverse transcriptase and DNA polymerase activities. This was carried out in the presence of oligonucleotide primers specific to regions of the HIV-1 gag genes.
Detection of HIV-1 RNA copies was done using an ELISA system. Denatured amplicons were bound to microwells serving as a solid support, then avidin horseradish peroxidase was added to the microwells. The quantity of amplicons were detected by adding a colorimetric substrate and reading was done using a spectrophotometer at 450 nm. This system uses a quantitation standard at a known concentration that is added to each sample before extraction and serves both to quantify the sample HIV-1 RNA and to compensate for plasma inhibitory factors affecting extraction and amplification.
Automated method using the COBAS Amplicor HIV-1 ultrasensitive method
Eighteen samples were also tested with the automated method. They were all run in one run on the same day. This was conducted according to the manufacturer’s instructions. The procedure was essentially identical to that of the manual method using Amplicor HIV-1 monitor, except that in this case, RT-PCR and detection were automatically carried out using a COBAS Amplicor analyser (Roche, Germany).
Statistical Analysis
Data were analyzed using the SAS version 8.2 software by the SAS Institute Inc (Cary, North Carolina). Descriptive statistics (means and standard deviations for continuous variables, frequencies for categorical variables) were used to summarize the data. Because of the relatively small sample size, a Spearman rank correlation coefficient was used to assess the relation between manual and automated viral load measures.
Results
Eighteen patients (9 males and 9 females) had viral loads determined by manual and automated methods. In these 18 patients, manual viral load values ranged from <50 to 220 000 copies/mL, with a median of 47 297, while automated viral load values ranged from 120 to 420 000 copies/mL with a median of 37 000 copies/mL (Table 1 ). Both methods were highly correlated, with a Spearman rank correlation statistic of .84, P < .0001 (Figure 1 ). Also, they were not statistically different from each other (Wilcoxon rank-sum statistic 532.5, t-approximation P value .9866).
Comparison of Viral Loads (Copies/mL) as Determined by Automated and Manual Methods in 18 HIV-Infected Patients in Yaounde, Cameroon
a The minimum viral load detected by the manual method was 50 copies/mL, a value of 25 (the mid-value) was used in calculating means and standard deviations.
Correlation between viral load (log copies/mL) as determined by automated and manual methods in 18 HIV-infected patients in Yaounde, Cameroon. A linear trend line is superimposed.
Discussion
This study aimed at evaluating the performance of a manual method of viral load detection with the Amplicor HIV-1 Monitor as could be used in day-to-day practice in a resource-limited setting, with high HIV genetic diversity. The manual method and the automated method were highly correlated.
This finding of a high correlation between an automated and a manual method was consistent with previous reports from other developed and resource-limited settings. This correlation was greater than 94% in comparisons between the automated Roche TaqMan real-time quantitative HIV-1 RNA PCR assay and the Roche AMPLICOR Version 1.5 conventional PCR assay 2 ; the Cobas AmpliPrep/Cobas Amplicor HIV-1 Monitor Ultrasensitive Test and the Cobas Amplicor HIV-1 Monitor test with manual specimen preparation 3 ; the Cobas AmpliPrep/Cobas Amplicor HIV Monitor Ultrasensitive Test (with automated specimen preparation) and the Cobas Amplicor HIV-1 Monitor Ultrasensitive Test (with manual sample preparation). 4 A lower correlation (85.6%) was found in comparing the automated Roche TaqMan real-time quantitative HIV-1 RNA PCR assay and the Roche AMPLICOR Monitor Version 1.5 in Ivory Coast, 5 suggesting the need for these methods to be assessed in each setting, prior to their use.
It is recommended that CD4 counts be supplemented by viral load counts in the evaluation of response to antiretroviral therapy. The possibility of using a cheaper alternative would greatly enhance the ability to implement this recommendation. The cost of equipment required to start running the COBAS method is generally more expensive than that of the manual method. The equipment used for running the manual viral load (thermal cycler, microplate washer and reader, incubator) can also be used for other diagnostic procedures (other than viral load detection) and will further reduce the relative cost of the manual test. On the other hand, unlike the manual method, with COBAS Amplicor, amplification through detection procedures do not occupy the technician who can therefore concentrate on other duties. Contrary to the manual method that is exposed to human errors and more labor (from repeated reagent add and wash cycles), the automated nature of the COBAS manipulation also implies a limitation in errors in the amplification to detection stages of the analysis. Despite the time, attention requirements, and potential for errors, the manual method when correctly operated gives results comparable to those of the automated method. It therefore appears adequate to monitor patients in resource-poor settings such as ours.
While this study is the first to report such data in Cameroon, the major limitations were the relatively small sample size and the lack of data on HIV clades. Further studies assessing the impact of HIV clade diversity in the performance of these tests as well as a detailed cost-effectiveness analysis would be of particular importance in guiding the choice of what assay to use in this setting.
Conclusions
In conclusion, this study, which aimed at comparing the manual and automated methods, showed results of manual viral load that correlated with results of the COBAS Amplicor. Even though it demands more effort from the laboratory technician than the COBAS Amplicor, the manual method appears adequate to use in monitoring HIV patients on treatment in resource-limited settings.
Footnotes
The author(s) declared no conflicts of interest with respect to the authorship and/or publication of this article.
Acknowledgements
We thank all study participants. The CSCCD where this study was conducted was supported by a training grant from the European Union. JA was a Fogarty AITRP fellow, supported by NIH Fogarty grant DHHS/NIH/FIC 5 D43 TW01039-08 AIDS International Training and Research Program to the University of North Carolina at Chapel Hill.
The author(s) received no financial support for the research and/or authorship of this article.