Abstract
In the Roche Amplicor HIV-1 Monitor Test, the ratio of the optical density used to estimate the RNA concentration and the optical density of the preceding or the following dilution can be used to identify specimens with inaccurate results which should be retested. We present an algorithm for the identification of such results.
Measurements of human immunodeficiency virus (HIV) RNA concentrations are widely used to inform treatment decisions for patients with HIV infection (5-11). Aspects of systematic and random variations in RNA measurements have been examined in several studies (1-4, 12), but the translation of these studies into correlates for assessments of the outcomes of individual assays is often difficult. Here we describe the use of a ratio of successive optical densities (ODs) from the Roche HIV-1 Monitor Test to assess the accuracies of individual assay results.
The algorithm used to estimate the HIV RNA concentration in the Monitor assay motivated this analysis. This algorithm is based on three assumptions: (i) the ODs from the serial fivefold dilutions of the PCR-amplified product form a pattern of decreasing ODs; (ii) the OD varies linearly with the HIV RNA concentration, at least for ODs between 0.20 and 2.0 units; and (iii) the intercept of the linear relationship is zero after the background of 0.07 units is subtracted from the OD. These assumptions imply that the ratio of the larger to the smaller of two ODs that fall in the linear range should be close to 5. Therefore, departures from OD ratios of 5 could signal problems with the accuracies of assay results.
This idea was tested by using data collected as part of the Virology Quality Assessment (VQA) Program. The VQA Program was established by the Division of AIDS (DAIDS) of the National Institute of Allergy and Infectious Diseases, National Institutes of Health, to provide quality assurance for clinical trials and other DAIDS-sponsored studies (12). As part of the VQA Program, coded test panels and RNA copy controls made from a well-characterized HIV type 1 (HIV-1) isolate spiked into human plasma at known concentrations (50 to 750,000 copies/ml) were periodically sent to participating laboratories (January 1997 to February 2001) to assess laboratory proficiency in performing quantitative HIV RNA assays. A total of 14,069 results from the Standard Roche Monitor Test and 3,833 results from the UltraSensitive Monitor Test were examined. The results obtained with the two versions of the kit were very similar, so they were combined for this analysis.
In 12,846 (71.8%) assays, two or more consecutive ODs fell in the range used for estimation of the HIV RNA concentration (0.20 to 2.0 units). Ratios were formed by using the last two ODs in this range, after subtracting the background of 0.07 units from both (i.e., the OD used to estimate the HIV RNA concentration formed the denominator of the OD ratio). The results for assays with two to five ODs between 0.20 and 2.0 were very similar, so they were pooled. The median recovery ratio (estimated RNA concentration divided by nominal RNA concentration) for these assays was close to 1.0 (100%) when the OD ratio was close to 5.0 (Table 1). The recovery ratio varied inversely with the OD ratio. The estimated HIV RNA concentration differed from the nominal concentration by greater than twofold (i.e., recovery ratio <0.5 or >2) in 1,834 (14.3%) assays. Departures of this magnitude were more common at relatively high or low OD ratios than they were at OD ratios near 5.0. At OD ratios <3 or >7, 38% of the recovery ratios were <0.50 or >2.0, but at OD ratios between 3 and 7, only 13% of recovery ratios had errors of twofold or greater.
TABLE 1.
Relationship between OD ratio and recovery ratio in the Standard and UltraSensitive Roche Monitor assays
| Recovery ratio | No. (%) of samples
|
|||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| OD ratio involving two OD values between 0.2 and 2.0
|
OD ratio involving one OD value between 0.2 and 2.0 and one OD value >2.0
|
OD ratio involving one OD value between 0.2 and 2.0 and one OD value <0.2
|
||||||||||
| <3.0 | 3.0-7.0 | ≥7.0 | Total | <3.0 | 3.0-7.0 | ≥7.0 | Total | <3.0 | 3.0-7.0 | ≥7.0 | Total | |
| <0.50 | 19 (3.9) | 835 (6.8) | 24 (32.9) | 878 (6.8) | 6 (9.1) | 266 (7.5) | 6 (42.9) | 278 (7.9) | 6 (21.4) | 318 (8.3) | 215 (18.3) | 539 (10.7) |
| ≥0.50, <2.0 | 299 (61.8) | 10.666 (86.8) | 47 (64.4) | 11,012 (85.7) | 56 (84.8) | 3,144 (88.1) | 8 (57.1) | 3,208 (87.9) | 17 (60.7) | 3,352 (87.0) | 910 (77.3) | 4,279 (84.6) |
| ≥2.0 | 166 (34.3) | 788 (6.4) | 2 (2.7) | 956 (7.4) | 4 (6.1) | 158 (4.4) | 0 (0.0) | 162 (4.4) | 5 (17.9) | 181 (4.7) | 52 (4.4) | 238 (4.7) |
| Total | 484 | 12,289 | 73 | 12,846 | 66 | 3,568 | 14 | 3,648 | 28 | 3,851 | 1,177 | 5,056 |
In the remaining 5,056 assays only one OD fell between 0.20 and 2.0 units, so ratios for these assays involved one OD that was either <0.20 or >2.0. In the 3,648 assays with a single OD between 0.2 and 2.0 (which occurred in the second through the sixth dilutions), OD ratios were formed by using that OD and the OD at the preceding dilution. Estimates were biased in 440 (12.1%) assays; the result was biased in 43% of the assays with OD ratios >7 (6 assays), but were biased in 15% of the assays with OD ratios <3 (10 assays) (Table 1).
Ratios involving the OD between 0.2 and 2.0 and the OD at the next dilution (i.e., OD < 0.2) were calculated for all 5,056 assays. A biased result was obtained in 777 (15.4%) assays; 39.3% of the assays with OD ratios <3 were biased, and 22.7% of the assays with OD ratios >7 were biased.
The intent of this study was to identify an algorithm that could be used to identify potentially biased HIV-1 RNA results that would probably go undetected even if the manufacturer's guidelines for the assay were strictly followed. Seventy-two percent of RNA assays evaluated contained two or more ODs between 0.2 and 2.0. Of these assays, 38% were inaccurate if the OD ratios were not between 3 and 7. This suggests that these assays should be repeated. Since the error is associated with the detection phase of the assay, a simple redetection should correct the problem. In assays in which only one wild-type OD value is between 0.2 and 2.0, the OD ratio should be formed by using the OD value used to calculate the RNA concentration estimate and the OD following it. If the OD ratio is not between 3 and 7, then this assay should be repeated as well.
All of the results presented here were obtained with microwell plate version 1.0 of the Monitor assay. These rules cannot be applied to the COBAS version of this assay because a different dilution scheme and algorithm are used. Extreme OD ratios may occur less frequently with the COBAS version, which uses automated amplification and detection, but this will require further investigation.
Acknowledgments
This work was supported by NIAID contracts AI-35712 and AI-85354.
We thank William A. Meyer III for careful review of the manuscript.
REFERENCES
- 1.Bartlett, J. A., R. DeMasi, D. Dawson, and A. Hill. 1998. Variability in repeated consecutive measurements of plasma human immunodeficiency virus RNA in persons receiving stable nucleoside reverse transcriptase inhibitor therapy or no treatment. J. Infect. Dis. 178:1803-1805. [DOI] [PubMed] [Google Scholar]
- 2.Brambilla, D., P. S. Reichelderfer, J. W. Bremer, D. E. Shapiro, R. C. Hershow, D. A. Katzenstein, S. M. Hammer, B. Jackson, A. C. Collier, R. S. Sperling, M. G. Fowler, R. W. Coombs, et al. 1999. The contribution of assay variation and biological variation to the total variability of plasma HIV-1 RNA measurements. AIDS 13:2269-2279. [DOI] [PubMed] [Google Scholar]
- 3.Brambilla, D. J., S. Granger, C. Jennings, and J. W. Bremer. 2001. A multi-site comparison of the Standard and UltraSensitive Roche HIV-1 Monitor assays. J. Clin. Microbiol. 39:1121-1123. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Brambilla, D. J., S. Granger, C. Jennings, and J. W. Bremer. 2001. The effects of errors in the sequence of optical densities from the Roche HIV-1 Monitor assay on the validity of quantitative RNA estimates. J. Clin. Microbiol. 39:1118-1120. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hughes, M. D., V. A. Johnson, M. S. Hirsch, J. W. Bremer, T. Elbeik, A. Erice, D. R. Kuritzkes, W. A. Scott, S. A. Spector, N. Basgoz, M. A. Fischl, R. T. D'Aquila, et al. 1997. Monitoring plasma HIV-1 RNA levels in addition to CD4+ lymphocyte count improves assessment of antiretroviral therapeutic response. Ann. Intern. Med. 126:929-938. [DOI] [PubMed] [Google Scholar]
- 6.Katzenstein, D. A., S. M. Hammer, M. D. Hughes, H. Bundacker, J. B. Jackson, S. Fiscus, S. Rasheed, T. Elbeik, R. Reichman, A. Japour, T. C. Merigan, and M. S. Hirsch for the AIDS Clinical Trials Group Study 175 Virology Study Team. 1996. The relation of virologic and immunologic markers to clinical outcomes after nucleoside therapy in HIV-infected adults with 200 to 500 CD4 cells per cubic millimeter. N. Engl. J. Med. 335:1091-1098. [DOI] [PubMed] [Google Scholar]
- 7.Lathey, J. L., M. D. Hughes, S. A. Fiscus, T. Pi, J. B. Jackson, S. Rasheed, T. Elbeik, R. Reichman, A. Japour, R. T. D'Aquila, W. Scott, B. P. Griffith, S. M. Hammer, D. A. Katzenstein, et al. 1998. Variability and prognostic values of virologic and CD4 cell measures in human immunodeficiency virus type 1-infected patients with 200-500 CD4 cells/mm3 (ACTG 175). J. Infect. Dis. 177:617-624. [DOI] [PubMed] [Google Scholar]
- 8.Mellors, J. W., L. A. Kingsley, C. R. Rinaldo, Jr., J. A. Todd, B. S. Hoo, R. Kokka, and P. Gupta. 1995. Quantitation of HIV-1 RNA in plasma predicts outcome after seroconversion. Ann. Intern. Med. 122:573-579. [DOI] [PubMed] [Google Scholar]
- 9.Mellors, J. W., C. R. Rinaldo, P. Gupta, R. M. White, J. A. Todd, and L. A. Kingsley. 1996. Prognosis of HIV-1 infection predicted by the quantity of virus in plasma. Science 272:1167-1170. [DOI] [PubMed] [Google Scholar]
- 10.Saag, M. S. 1997. Use of HIV viral load in clinical practice: back to the future. Ann. Intern. Med. 126:983-985. [DOI] [PubMed] [Google Scholar]
- 11.Schooley, R. T. 1995. Correlation between viral load measurements and outcomes in clinical trials of antiviral drugs. AIDS 9(Suppl. 2):515-519. [PubMed] [Google Scholar]
- 12.Yen-Lieberman, B., D. Brambilla, B. Jackson, J. Bremer, R. Coombs, M. Cronin, S. Herman, D. Katzenstein, S. Leung, H. J. Lin, P. Palumbo, S. Rasheed, J. Todd, M. Vahey, and P. Reichelderfer. 1996. Evaluation of a quality assurance program for quantitation of human immunodeficiency virus type 1 RNA in plasma by the AIDS Clinical Trials Group virology laboratories. J. Clin. Microbiol. 34:2695-2701. [DOI] [PMC free article] [PubMed] [Google Scholar]