- Email: [email protected]
antibody assays using the p24 WHO standard
Journal of Clinical Virology 50 (2011) 57–60
Contents lists available at ScienceDirect
Journal of Clinical Virology journal homepage: www.elsevier.com/locate/jcv
Short communication
Analytical sensitivity of four HIV combined antigen/antibody assays using the p24 WHO standard Marcel Miedouge a,∗ , Monique Grèze b , Agnès Bailly b , Jacques Izopet a,c a
Laboratoire de Virologie, Institut Fédératif de Biologie de Purpan, 330 avenue de Grande Bretagne, TSA 40031, 31059 Toulouse, Cédex 9, France Laboratoire du Centre Hospitalier, 22 boulevard Sibille, 81013, Albi Cédex 9, France c Inserm U563, Centre de Physiopathologie de Toulouse Purpan, 31024 Toulouse, Cédex 3, France b
a r t i c l e
i n f o
Article history: Received 28 July 2010 Received in revised form 7 September 2010 Accepted 8 September 2010 Keywords: HIV p24 antigen Serological window WHO standard
a b s t r a c t Background: The Common Technical Specifications for HIV-1 p24 assays published in 2009 fixed the lower limit of detection for obtaining C.E. approval at 2 IU/ml against the WHO standard (first international reference, code NIBSC 90/636); it was previously 50 pg/ml. Some recent fourth generation HIV assays that simultaneously detect antigen and antibody are equivalent to p24 assays, but they were mainly evaluated using p24 antigen standards in pg/ml and little is known of their performance with the IU/ml standard. Objectives: To evaluate four of the combined serological assays most commonly used for HIV diagnosis in France against the WHO standard in IU/ml. Study design: The analytical sensitivity of four combined p24 antigen and antibody assays (ARCHITECT HIV Ag/Ab Combo, AxSYM HIV Ag/Ab Combo, VIDAS HIV DUO Quick and VIDAS HIV DUO Ultra) and of one p24 assay (VIDAS HIV p24 II) were determined using dilutions of the WHO standard. Results: Four of the five assays had a lower limit of detection below 2 IU/ml: 1.24 for ARCHITECT Combo, 0.66 for VIDAS DUO Ultra, 0.43 for VIDAS DUO Quick and 0.73 to 1.15 for VIDAS p24, while that of AxSYM was close to 2 (1.94–2.25). Conclusions: We have provided the first data on the lower limit of detection of HIV combined assays using the IU/ml WHO standard and demonstrated the need for a single international standard for comparing assays. We recommend the use of this approach in medical laboratory to validate on site their methods. © 2010 Elsevier B.V. All rights reserved.
1. Background The serological window for detecting anti-HIV antibodies is approximately three weeks. HIV RNA detection may reduce this delay by 12 days and p24 antigen detection by 5 days.1 The Common Technical Specifications (CTS) of the European Union published in 20022 required p24 Ag assays to have a lower limit of detection (LLD) below 50 pg/ml against the AFSSAPS (Agence Franc¸aise de Sécurité SAnitaire des Produits de Santé) HIV-1 antigen standard (HIV-1 group M, subtype B) in order to obtain C.E. approval. In the last version published in 2009 this LLD was fixed at 2 IU/ml against the WHO standard (HIV-1 isolate, code NIBSC 90/636).3 Recent fourth generation HIV assays that detect both antigen and
antibody fulfil the requirements for the p24 assay.4 French legislation on HIV screening changed in June 2010. It now states that HIV screening must be performed using a combined antigen/antibody test that can detect at least 2 IU/ml of p24 antigen. Previous studies have used various standards (AFSSAPS standard provided by the French Society of Blood Transfusion, HIV p24 antigen standard from Abbott Laboratories) and this has led to confusion (Table 1).4–9 The AFSSAPS standard is expressed as a p24 antigen concentration or as an HIV antigen concentration (1 pg/ml of p24 antigen = 3.65 pg/ml of HIV antigen). The Abbott standard is expressed as p24 antigen, and 1 pg/ml = 5 pg/ml of the AFSSAPS standard.5 2. Objectives
Abbreviations: LLD, lower limit of detection; CTS, Common Technical Specifications; S/CO, signal to cut-off ratio; AFSSAPS, Agence Franc¸aise de Sécurité SAnitaire des Produits de Santé. ∗ Corresponding author. Tel.: +33 5 67 69 04 21; fax: +33 5 67 69 04 25. E-mail address: [email protected] (M. Miedouge). 1386-6532/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jcv.2010.09.003
Little is known about the performance of assays in terms of the WHO standard, or about the relationship between the 50 pg/ml and 2 IU/ml criteria of the new CTS. We have therefore evaluated the performance of five of the serological assays most commonly used for HIV diagnosis in France against this WHO standard.
58
M. Miedouge et al. / Journal of Clinical Virology 50 (2011) 57–60
Table 1 Overview of published lower limit of detection (LLD) of five HIV assays. Assay
Publication
ARCHITECT HIV Ag/Ac Combo
Kwon et al.9 Ly et al.4
LLD expressed in pg/ml of the corresponding standard AFSSAPS/SFTS HIV antigen
AxSYM HIV Ag/Ac Combo
5
Ly et al. Ly et al.7 Sickinger et al.8 Kwon et al.9 5
ABBOTT HIV p24 antigen <5
18 Above 25 4.4–10.2 14.5–21.2 10
Ly et al. Ly et al.4
Close to 25 11.5
VIDAS HIV Duo Quick
Ly et al.4
13
VIDAS HIV p24 II
Laperche et al.6
11.2
VIDAS HIV Duo Ultra
AFSSAPS/SFTS HIV p24 antigen
3. Study design
3.3. Analytical sensitivity
3.1. HIV assays
The five assays were compared using two-fold dilutions of the WHO standard from 8 to 0.0625 IU/ml, tested in duplicate. The LLD was calculated using a statistical regression (y = ax + b) where the limit is the positive threshold of each assay. The reproducibility of dilution was determined by testing three independent dilutions giving a concentration of 2 IU/ml.
3.1.1. ARCHITECT HIV Ag/Ab Combo The ARCHITECT HIV Ag/Ab Combo (Abbott Diagnostics, France) is a chemiluminescent magnetic microparticle-based immunoassay used to detect HIV-1 p24 antigen and antibody to HIV-1/2 with an automated, random access instrument.9,10 Specimens with signal to cut-off ratios (S/CO) of 1.00 or above are considered reactive.
3.1.2. AxSYM HIV Ag/Ab Combo The AxSYM HIV Ag/Ab Combo (Abbott Diagnostics, France), is a fluorescent microparticle enzyme immunoassay that detects HIV p24 antigen and antibody to HIV-1/2 with an automated, random access instrument.8,10 Specimens with S/CO of 1.00 or greater are considered reactive, with a gray zone between 0.90 and 1.0.
3.1.3. VIDAS HIV DUO Quick The VIDAS HIV DUO Quick (bioMerieux, France) is an enzymelinked fluorescent assay which permits the simultaneous detection of p24 antigen and antibody to HIV-1/2.4 Specimens with S/CO of 0.25 or greater are considered reactive.
3.1.4. VIDAS HIV DUO Ultra The VIDAS HIV DUO Ultra is similar to the VIDAS HIV DUO Quick but it has a longer run-time (120 min) and gives separate results for antigen and antibody. 11–13
3.4. Pre-seroconversion specimen We tested the serum obtained from a patient at the preseroconversion phase of HIV infection with an HIV-1 group M, subtype G strain. The serum was diluted with HIV negative serum to a concentration near the LLD of the Architect assay and then tested with all the assays in duplicate. 4. Results 4.1. Analytical sensitivity Tables 2 and 3 summarizes the LLD for p24 antigen obtained for the five assays after linear regression using the eight two-fold dilutions. The linearity was excellent with coefficient regressions above 0.9999. The LLDs were below 2 IU/ml for all the assays except the AxSYM Combo assay, for which it was close to 2 IU/ml. The means and coefficients of variation obtained for three independent dilutions of the standard (at 2 IU/ml) are shown in Table 4. 4.2. Pre-seroconversion specimen
3.1.5. VIDAS HIV p24 II The VIDAS HIV p24 II (bioMerieux, France) is an automated quantitative enzyme-linked fluorescent immunoassay for determining HIV-1 p24 antigen.14 Specimens with HIV antigen concentrations of 18.2 pg/ml and above are considered to be reactive, with a gray zone between 10.9 and 18.2.
The S/Co results for the pre-seroconversion specimen were 2.23 with the ARCHITECT Combo, 1.14 with the AxSYM Combo, 0.77 with the VIDAS DUO Ultra, and 0.68 with the VIDAS DUO Quick. The VIDAS HIV p24 gave a concentration of 31.4 pg/ml. We used the above regression linear equation to obtain the corresponding values in IU/ml: 2.8 for ARCHITECT Combo, 2.7 for AxSYM Combo, 2.5 for VIDAS DUO Ultra, 2.4 for VIDAS DUO Quick, and 1.9 for VIDAS p24.
3.2. p24 antigen standard
5. Discussion
The WHO International Standard HIV-1 p24 antigen (code 90/636) was provided by the National Institute for Biological Standards and Control (NISBC, Hertfordshire, UK, [email protected]). It was reconstituted in 1.0 ml of distilled water, aliquoted, and stored at −20 ◦ C. Working dilutions were made with HIV negative serum and tested with each assay within 24 h.
To our knowledge, only one study has used the WHO standard to compare assay and determine their LLDs.4 Their data are surprising since one assay that had an LLD over 50 pg/ml (126.5) was clearly below 2 IU/ml (0.373), which does not agree with the evolution of the CTS requirements.2,3 Four of the five assays test by us had LLDs below 2 IU/ml: 1.24 for ARCHITECT Combo, 0.66 for VIDAS DUO Ultra, 0.43 for VIDAS
M. Miedouge et al. / Journal of Clinical Virology 50 (2011) 57–60
59
Table 2 Lower limit of detection (LLD) of the five HIV assays against the WHO standard. NIBSC, P24 Ag IU/mL
8 4 2 1 0.5 0.25 0.125 0.0625 Linear regression parameters a
Reactivity threshold LLD IU/mL
Resultsa ARCHITECT HIV Ag/Ac Combo
AxSYM HIV Ag/Ac Combo
VIDAS HIV Duo Ultra
VIDAS HIV Duo Quick
VIDAS HIV p24 II
6.29 3.08 1.59 0.83 0.45 0.26 0.18 0.14
2.83 1.57 0.89 0.60 0.48 0.37 0.33 0.34
2.36 1.245 0.64 0.33 0.17 0.10 0.05 0.02
1.90 1.03 0.60 0.38 0.26 0.22 0.19 0.17
138.3 68.8 33.0 15.5 <10.9 <10.9 <10.9 <10.9
y = 1.28x − 0.046 R2 = 0.9997 1 1.24
y = 3.14x − 0.89 R2 = 0.9996 0.9–1 1.94–2.25
y = 3.46x − 0.20 R2 = 0.9994 0.25 0.66
y = 4.57x − 0.71 R2 = 0.9999 0.25 0.43
y = 0.057x + 0.11 R2 = 0.9999 10.9–18.2 0.73–1.15
a Results and reactivity thresholds correspond to the S/Co ratios except for the VIDAS HIV p24 II that is expressed in pg/ml. For each dilution we reported the mean of the duplicates.
Table 3 Comparison of available data for lower limit of detection (LLD) of p24 antigen in pg/ml for the five HIV assays versus our results expressed in IU/ml. Assay
ARCHITECT HIV Ag/Ac Combo AxSYM HIV Ag/Ac Combo VIDAS HIV Duo Ultra VIDAS HIV Duo Quick VIDAS HIV p24 II
LLD depending on the corresponding standard Published results against AFSSAPS HIV Antigen standard (pg/ml)
Manufacturer data against AFSSAPS HIV Antigen standard (pg/ml)
Our results against the WHO standard (IU/ml)
18 to <25 22–77.4 11.5–25 13 11.2
18.1 <91.25 11.5 16.5 11.25
1.24 1.94–2.25 0.66 0.43 0.73–1.15
Table 4 Inter-assay variation determined with 3 independent dilutions at 2 IU/ml of the WHO p24 antigen standard. WHO standard 2 IU/ml
Dilution 1 Dilution 2 Dilution 3 Mean Coefficient of variation %
Results (expressed as S/Co ratios) Architect
Axsym
Vidas Ultra
Vidas Quick
Vidas p24 II
1.59 1.54 1.43 1.52 5.4
0.89 0.92 0.91 0.90 1.5
0.64 0.57 0.52 0.58 10.2
0.6 0.59 0.55 0.58 4.8
33.03 29.93 27.62 30.19 9.0
DUO Quick and 0.73–1.15 for VIDAS p24, while the LLD for AxSYM was close to 2 (1.94–2.25). Our results are not in agreement with those of Ly et al.,4 who found LLDs of 0.0677 for ARCHITECT Combo, 0.0426 for VIDAS DUO Ultra, and 0.0539 IU/ml for VIDAS DUO Quick, probably because the standard was from Boston Biomedica, while ours was from NIBSC. If the LLDs obtained in published studies or given in the assay instructions are expressed using the same reference (AFFSAPS standard in pg/ml of HIV antigen), our results using the WHO standard appear to be quite concordant (Table 3). The ARCHITECT Combo, VIDAS DUO Ultra, VIDAS DUO Quick and VIDAS p24 assays all satisfy the CTS requirements regardless of the standard used, while the AxSYM combo is at the limit of this specification. More extensive data are now needed to confirm these results because many factors can generate uncertainty. First is the uncertainty in the preparation of the NIBSC standard, which is not available. Second, the dilution process and the inter-assay variation may also interfere: we obtained a 1.5–10.2 coefficient of variation, which would probably be greater if different lots of reagents were used. It is also important to compare assays using seroconversion panels and various strains of HIV. The test we did with a serum from a patient infected with a HIV-1 subtype G strain that was diluted to obtain a concentration close to 2 IU/ml gave concordant results with all the systems tested, although there was a small difference between the combined tests (2.4–2.8) and the p24 assay (1.9). This
is not surprising since these tests are not standardized against the WHO standard. Further studies are needed to establish the relationship between of p24 antigen IU/ml and HIV RNA copies/ml. To conclude, we have produced a set of data showing the correspondence between pg/ml standards and the IU/ml WHO standard, and demonstrated the need for a single international standard for comparing assays. We also recommend this approach in medical laboratories to validate on site their methods. Conflict of interest None. References 1. Fiebig EW, Wright DJ, Rawal BD, Garrett PE, Schumacher RT, Peddada L, et al. Dynamics of HIV viremia and antibody seroconversion in plasma donors: implications for diagnosis and staging of primary HIV infection. AIDS 2003;17(13):1871–9. 2. Commission decision on common technical specifications for in vitro diagnostic medical devices. Off J Eur Commun 2002;L131:17–30. 3. Commission decision on common technical specifications for in vitro diagnostic medical devices. Off J Eur Commun 2009;L39:3–49. 4. Ly TD, Ebel A, Faucher V, Fihman V, Laperche S. Could the new HIV combined p24 antigen and antibody assays replace p24 antigen specific assays? J Virol Methods 2007;143(1):86–94. 5. Ly TD, Martin L, Daghfal D, Sandridge A, West D, Bristow R, et al. Seven human immunodeficiency virus (HIV) antigen-antibody combination assays: evalua-
60
6. 7.
8.
9.
M. Miedouge et al. / Journal of Clinical Virology 50 (2011) 57–60 tion of HIV seroconversion sensitivity and subtype detection. J Clin Microbiol 2001;39(9):3122–8. Laperche S, Ly TD. Sensitivity of HIV infection screening assays in 2001. Ann Biol Clin (Paris) 2002;60(3):307–15. Ly TD, Laperche S, Brennan C, Vallari A, Ebel A, Hunt J, et al. Evaluation of the sensitivity and specificity of six HIV combined p24 antigen and antibody assays. J Virol Methods 2004;122(2):185–94. Sickinger E, Stieler M, Kaufman B, Kapprell HP, West D, Sandridge A, et al. AxSYM Clinical Study Group Multicenter evaluation of a new, automated enzyme-linked immunoassay for detection of human immunodeficiency virus-specific antibodies and antigen. J Clin Microbiol 2004;42(1): 21–9. Kwon JA, Yoon SY, Lee CK, Lim CS, Lee KN, Sung HJ, et al. Performance evaluation of three automated human immunodeficiency virus antigen-antibody combination immunoassays. J Virol Methods 2006;133(1):20–6.
10. Malm K, von Sydow M, Andersson S. Performance of three automated fourthgeneration combined HIV antigen/antibody assays in large-scale screening of blood donors and clinical samples. Transfus Med 2009;19(2):78–88. 11. Saville RD, Constantine NT, Cleghorn FR, Jack N, Bartholomew C, Edwards J, et al. Fourth-generation enzyme-linked immunosorbent assay for the simultaneous detection of human immunodeficiency virus antigen and antibody. J Clin Microbiol 2001;39(7):2518–24. 12. Weber B, Berger A, Rabenau H, Doerr HW. Evaluation of a new combined antigen and antibody human immunodeficiency virus screening assay VIDAS HIV DUO Ultra. J Clin Microbiol 2002;40(4):1420–6. 13. Bourlet T, Pretis C, Pillet S, Lesenechal M, Piche J, Pozzetto B. Comparative evaluation of the VIDAS HIV DUO Ultra assay for combined detection of HIV-1 antigen and antibodies to HIV. J Virol Methods 2005;127(2):165–7. 14. Fransen K, Beelaert G, van der Groen G. Evaluation of four HIV antigen tests. J Virol Methods 2001;93(1–2):189–93.
Contents lists available at ScienceDirect
Journal of Clinical Virology journal homepage: www.elsevier.com/locate/jcv
Short communication
Analytical sensitivity of four HIV combined antigen/antibody assays using the p24 WHO standard Marcel Miedouge a,∗ , Monique Grèze b , Agnès Bailly b , Jacques Izopet a,c a
Laboratoire de Virologie, Institut Fédératif de Biologie de Purpan, 330 avenue de Grande Bretagne, TSA 40031, 31059 Toulouse, Cédex 9, France Laboratoire du Centre Hospitalier, 22 boulevard Sibille, 81013, Albi Cédex 9, France c Inserm U563, Centre de Physiopathologie de Toulouse Purpan, 31024 Toulouse, Cédex 3, France b
a r t i c l e
i n f o
Article history: Received 28 July 2010 Received in revised form 7 September 2010 Accepted 8 September 2010 Keywords: HIV p24 antigen Serological window WHO standard
a b s t r a c t Background: The Common Technical Specifications for HIV-1 p24 assays published in 2009 fixed the lower limit of detection for obtaining C.E. approval at 2 IU/ml against the WHO standard (first international reference, code NIBSC 90/636); it was previously 50 pg/ml. Some recent fourth generation HIV assays that simultaneously detect antigen and antibody are equivalent to p24 assays, but they were mainly evaluated using p24 antigen standards in pg/ml and little is known of their performance with the IU/ml standard. Objectives: To evaluate four of the combined serological assays most commonly used for HIV diagnosis in France against the WHO standard in IU/ml. Study design: The analytical sensitivity of four combined p24 antigen and antibody assays (ARCHITECT HIV Ag/Ab Combo, AxSYM HIV Ag/Ab Combo, VIDAS HIV DUO Quick and VIDAS HIV DUO Ultra) and of one p24 assay (VIDAS HIV p24 II) were determined using dilutions of the WHO standard. Results: Four of the five assays had a lower limit of detection below 2 IU/ml: 1.24 for ARCHITECT Combo, 0.66 for VIDAS DUO Ultra, 0.43 for VIDAS DUO Quick and 0.73 to 1.15 for VIDAS p24, while that of AxSYM was close to 2 (1.94–2.25). Conclusions: We have provided the first data on the lower limit of detection of HIV combined assays using the IU/ml WHO standard and demonstrated the need for a single international standard for comparing assays. We recommend the use of this approach in medical laboratory to validate on site their methods. © 2010 Elsevier B.V. All rights reserved.
1. Background The serological window for detecting anti-HIV antibodies is approximately three weeks. HIV RNA detection may reduce this delay by 12 days and p24 antigen detection by 5 days.1 The Common Technical Specifications (CTS) of the European Union published in 20022 required p24 Ag assays to have a lower limit of detection (LLD) below 50 pg/ml against the AFSSAPS (Agence Franc¸aise de Sécurité SAnitaire des Produits de Santé) HIV-1 antigen standard (HIV-1 group M, subtype B) in order to obtain C.E. approval. In the last version published in 2009 this LLD was fixed at 2 IU/ml against the WHO standard (HIV-1 isolate, code NIBSC 90/636).3 Recent fourth generation HIV assays that detect both antigen and
antibody fulfil the requirements for the p24 assay.4 French legislation on HIV screening changed in June 2010. It now states that HIV screening must be performed using a combined antigen/antibody test that can detect at least 2 IU/ml of p24 antigen. Previous studies have used various standards (AFSSAPS standard provided by the French Society of Blood Transfusion, HIV p24 antigen standard from Abbott Laboratories) and this has led to confusion (Table 1).4–9 The AFSSAPS standard is expressed as a p24 antigen concentration or as an HIV antigen concentration (1 pg/ml of p24 antigen = 3.65 pg/ml of HIV antigen). The Abbott standard is expressed as p24 antigen, and 1 pg/ml = 5 pg/ml of the AFSSAPS standard.5 2. Objectives
Abbreviations: LLD, lower limit of detection; CTS, Common Technical Specifications; S/CO, signal to cut-off ratio; AFSSAPS, Agence Franc¸aise de Sécurité SAnitaire des Produits de Santé. ∗ Corresponding author. Tel.: +33 5 67 69 04 21; fax: +33 5 67 69 04 25. E-mail address: [email protected] (M. Miedouge). 1386-6532/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jcv.2010.09.003
Little is known about the performance of assays in terms of the WHO standard, or about the relationship between the 50 pg/ml and 2 IU/ml criteria of the new CTS. We have therefore evaluated the performance of five of the serological assays most commonly used for HIV diagnosis in France against this WHO standard.
58
M. Miedouge et al. / Journal of Clinical Virology 50 (2011) 57–60
Table 1 Overview of published lower limit of detection (LLD) of five HIV assays. Assay
Publication
ARCHITECT HIV Ag/Ac Combo
Kwon et al.9 Ly et al.4
LLD expressed in pg/ml of the corresponding standard AFSSAPS/SFTS HIV antigen
AxSYM HIV Ag/Ac Combo
5
Ly et al. Ly et al.7 Sickinger et al.8 Kwon et al.9 5
ABBOTT HIV p24 antigen <5
18 Above 25 4.4–10.2 14.5–21.2 10
Ly et al. Ly et al.4
Close to 25 11.5
VIDAS HIV Duo Quick
Ly et al.4
13
VIDAS HIV p24 II
Laperche et al.6
11.2
VIDAS HIV Duo Ultra
AFSSAPS/SFTS HIV p24 antigen
3. Study design
3.3. Analytical sensitivity
3.1. HIV assays
The five assays were compared using two-fold dilutions of the WHO standard from 8 to 0.0625 IU/ml, tested in duplicate. The LLD was calculated using a statistical regression (y = ax + b) where the limit is the positive threshold of each assay. The reproducibility of dilution was determined by testing three independent dilutions giving a concentration of 2 IU/ml.
3.1.1. ARCHITECT HIV Ag/Ab Combo The ARCHITECT HIV Ag/Ab Combo (Abbott Diagnostics, France) is a chemiluminescent magnetic microparticle-based immunoassay used to detect HIV-1 p24 antigen and antibody to HIV-1/2 with an automated, random access instrument.9,10 Specimens with signal to cut-off ratios (S/CO) of 1.00 or above are considered reactive.
3.1.2. AxSYM HIV Ag/Ab Combo The AxSYM HIV Ag/Ab Combo (Abbott Diagnostics, France), is a fluorescent microparticle enzyme immunoassay that detects HIV p24 antigen and antibody to HIV-1/2 with an automated, random access instrument.8,10 Specimens with S/CO of 1.00 or greater are considered reactive, with a gray zone between 0.90 and 1.0.
3.1.3. VIDAS HIV DUO Quick The VIDAS HIV DUO Quick (bioMerieux, France) is an enzymelinked fluorescent assay which permits the simultaneous detection of p24 antigen and antibody to HIV-1/2.4 Specimens with S/CO of 0.25 or greater are considered reactive.
3.1.4. VIDAS HIV DUO Ultra The VIDAS HIV DUO Ultra is similar to the VIDAS HIV DUO Quick but it has a longer run-time (120 min) and gives separate results for antigen and antibody. 11–13
3.4. Pre-seroconversion specimen We tested the serum obtained from a patient at the preseroconversion phase of HIV infection with an HIV-1 group M, subtype G strain. The serum was diluted with HIV negative serum to a concentration near the LLD of the Architect assay and then tested with all the assays in duplicate. 4. Results 4.1. Analytical sensitivity Tables 2 and 3 summarizes the LLD for p24 antigen obtained for the five assays after linear regression using the eight two-fold dilutions. The linearity was excellent with coefficient regressions above 0.9999. The LLDs were below 2 IU/ml for all the assays except the AxSYM Combo assay, for which it was close to 2 IU/ml. The means and coefficients of variation obtained for three independent dilutions of the standard (at 2 IU/ml) are shown in Table 4. 4.2. Pre-seroconversion specimen
3.1.5. VIDAS HIV p24 II The VIDAS HIV p24 II (bioMerieux, France) is an automated quantitative enzyme-linked fluorescent immunoassay for determining HIV-1 p24 antigen.14 Specimens with HIV antigen concentrations of 18.2 pg/ml and above are considered to be reactive, with a gray zone between 10.9 and 18.2.
The S/Co results for the pre-seroconversion specimen were 2.23 with the ARCHITECT Combo, 1.14 with the AxSYM Combo, 0.77 with the VIDAS DUO Ultra, and 0.68 with the VIDAS DUO Quick. The VIDAS HIV p24 gave a concentration of 31.4 pg/ml. We used the above regression linear equation to obtain the corresponding values in IU/ml: 2.8 for ARCHITECT Combo, 2.7 for AxSYM Combo, 2.5 for VIDAS DUO Ultra, 2.4 for VIDAS DUO Quick, and 1.9 for VIDAS p24.
3.2. p24 antigen standard
5. Discussion
The WHO International Standard HIV-1 p24 antigen (code 90/636) was provided by the National Institute for Biological Standards and Control (NISBC, Hertfordshire, UK, [email protected]). It was reconstituted in 1.0 ml of distilled water, aliquoted, and stored at −20 ◦ C. Working dilutions were made with HIV negative serum and tested with each assay within 24 h.
To our knowledge, only one study has used the WHO standard to compare assay and determine their LLDs.4 Their data are surprising since one assay that had an LLD over 50 pg/ml (126.5) was clearly below 2 IU/ml (0.373), which does not agree with the evolution of the CTS requirements.2,3 Four of the five assays test by us had LLDs below 2 IU/ml: 1.24 for ARCHITECT Combo, 0.66 for VIDAS DUO Ultra, 0.43 for VIDAS
M. Miedouge et al. / Journal of Clinical Virology 50 (2011) 57–60
59
Table 2 Lower limit of detection (LLD) of the five HIV assays against the WHO standard. NIBSC, P24 Ag IU/mL
8 4 2 1 0.5 0.25 0.125 0.0625 Linear regression parameters a
Reactivity threshold LLD IU/mL
Resultsa ARCHITECT HIV Ag/Ac Combo
AxSYM HIV Ag/Ac Combo
VIDAS HIV Duo Ultra
VIDAS HIV Duo Quick
VIDAS HIV p24 II
6.29 3.08 1.59 0.83 0.45 0.26 0.18 0.14
2.83 1.57 0.89 0.60 0.48 0.37 0.33 0.34
2.36 1.245 0.64 0.33 0.17 0.10 0.05 0.02
1.90 1.03 0.60 0.38 0.26 0.22 0.19 0.17
138.3 68.8 33.0 15.5 <10.9 <10.9 <10.9 <10.9
y = 1.28x − 0.046 R2 = 0.9997 1 1.24
y = 3.14x − 0.89 R2 = 0.9996 0.9–1 1.94–2.25
y = 3.46x − 0.20 R2 = 0.9994 0.25 0.66
y = 4.57x − 0.71 R2 = 0.9999 0.25 0.43
y = 0.057x + 0.11 R2 = 0.9999 10.9–18.2 0.73–1.15
a Results and reactivity thresholds correspond to the S/Co ratios except for the VIDAS HIV p24 II that is expressed in pg/ml. For each dilution we reported the mean of the duplicates.
Table 3 Comparison of available data for lower limit of detection (LLD) of p24 antigen in pg/ml for the five HIV assays versus our results expressed in IU/ml. Assay
ARCHITECT HIV Ag/Ac Combo AxSYM HIV Ag/Ac Combo VIDAS HIV Duo Ultra VIDAS HIV Duo Quick VIDAS HIV p24 II
LLD depending on the corresponding standard Published results against AFSSAPS HIV Antigen standard (pg/ml)
Manufacturer data against AFSSAPS HIV Antigen standard (pg/ml)
Our results against the WHO standard (IU/ml)
18 to <25 22–77.4 11.5–25 13 11.2
18.1 <91.25 11.5 16.5 11.25
1.24 1.94–2.25 0.66 0.43 0.73–1.15
Table 4 Inter-assay variation determined with 3 independent dilutions at 2 IU/ml of the WHO p24 antigen standard. WHO standard 2 IU/ml
Dilution 1 Dilution 2 Dilution 3 Mean Coefficient of variation %
Results (expressed as S/Co ratios) Architect
Axsym
Vidas Ultra
Vidas Quick
Vidas p24 II
1.59 1.54 1.43 1.52 5.4
0.89 0.92 0.91 0.90 1.5
0.64 0.57 0.52 0.58 10.2
0.6 0.59 0.55 0.58 4.8
33.03 29.93 27.62 30.19 9.0
DUO Quick and 0.73–1.15 for VIDAS p24, while the LLD for AxSYM was close to 2 (1.94–2.25). Our results are not in agreement with those of Ly et al.,4 who found LLDs of 0.0677 for ARCHITECT Combo, 0.0426 for VIDAS DUO Ultra, and 0.0539 IU/ml for VIDAS DUO Quick, probably because the standard was from Boston Biomedica, while ours was from NIBSC. If the LLDs obtained in published studies or given in the assay instructions are expressed using the same reference (AFFSAPS standard in pg/ml of HIV antigen), our results using the WHO standard appear to be quite concordant (Table 3). The ARCHITECT Combo, VIDAS DUO Ultra, VIDAS DUO Quick and VIDAS p24 assays all satisfy the CTS requirements regardless of the standard used, while the AxSYM combo is at the limit of this specification. More extensive data are now needed to confirm these results because many factors can generate uncertainty. First is the uncertainty in the preparation of the NIBSC standard, which is not available. Second, the dilution process and the inter-assay variation may also interfere: we obtained a 1.5–10.2 coefficient of variation, which would probably be greater if different lots of reagents were used. It is also important to compare assays using seroconversion panels and various strains of HIV. The test we did with a serum from a patient infected with a HIV-1 subtype G strain that was diluted to obtain a concentration close to 2 IU/ml gave concordant results with all the systems tested, although there was a small difference between the combined tests (2.4–2.8) and the p24 assay (1.9). This
is not surprising since these tests are not standardized against the WHO standard. Further studies are needed to establish the relationship between of p24 antigen IU/ml and HIV RNA copies/ml. To conclude, we have produced a set of data showing the correspondence between pg/ml standards and the IU/ml WHO standard, and demonstrated the need for a single international standard for comparing assays. We also recommend this approach in medical laboratories to validate on site their methods. Conflict of interest None. References 1. Fiebig EW, Wright DJ, Rawal BD, Garrett PE, Schumacher RT, Peddada L, et al. Dynamics of HIV viremia and antibody seroconversion in plasma donors: implications for diagnosis and staging of primary HIV infection. AIDS 2003;17(13):1871–9. 2. Commission decision on common technical specifications for in vitro diagnostic medical devices. Off J Eur Commun 2002;L131:17–30. 3. Commission decision on common technical specifications for in vitro diagnostic medical devices. Off J Eur Commun 2009;L39:3–49. 4. Ly TD, Ebel A, Faucher V, Fihman V, Laperche S. Could the new HIV combined p24 antigen and antibody assays replace p24 antigen specific assays? J Virol Methods 2007;143(1):86–94. 5. Ly TD, Martin L, Daghfal D, Sandridge A, West D, Bristow R, et al. Seven human immunodeficiency virus (HIV) antigen-antibody combination assays: evalua-
60
6. 7.
8.
9.
M. Miedouge et al. / Journal of Clinical Virology 50 (2011) 57–60 tion of HIV seroconversion sensitivity and subtype detection. J Clin Microbiol 2001;39(9):3122–8. Laperche S, Ly TD. Sensitivity of HIV infection screening assays in 2001. Ann Biol Clin (Paris) 2002;60(3):307–15. Ly TD, Laperche S, Brennan C, Vallari A, Ebel A, Hunt J, et al. Evaluation of the sensitivity and specificity of six HIV combined p24 antigen and antibody assays. J Virol Methods 2004;122(2):185–94. Sickinger E, Stieler M, Kaufman B, Kapprell HP, West D, Sandridge A, et al. AxSYM Clinical Study Group Multicenter evaluation of a new, automated enzyme-linked immunoassay for detection of human immunodeficiency virus-specific antibodies and antigen. J Clin Microbiol 2004;42(1): 21–9. Kwon JA, Yoon SY, Lee CK, Lim CS, Lee KN, Sung HJ, et al. Performance evaluation of three automated human immunodeficiency virus antigen-antibody combination immunoassays. J Virol Methods 2006;133(1):20–6.
10. Malm K, von Sydow M, Andersson S. Performance of three automated fourthgeneration combined HIV antigen/antibody assays in large-scale screening of blood donors and clinical samples. Transfus Med 2009;19(2):78–88. 11. Saville RD, Constantine NT, Cleghorn FR, Jack N, Bartholomew C, Edwards J, et al. Fourth-generation enzyme-linked immunosorbent assay for the simultaneous detection of human immunodeficiency virus antigen and antibody. J Clin Microbiol 2001;39(7):2518–24. 12. Weber B, Berger A, Rabenau H, Doerr HW. Evaluation of a new combined antigen and antibody human immunodeficiency virus screening assay VIDAS HIV DUO Ultra. J Clin Microbiol 2002;40(4):1420–6. 13. Bourlet T, Pretis C, Pillet S, Lesenechal M, Piche J, Pozzetto B. Comparative evaluation of the VIDAS HIV DUO Ultra assay for combined detection of HIV-1 antigen and antibodies to HIV. J Virol Methods 2005;127(2):165–7. 14. Fransen K, Beelaert G, van der Groen G. Evaluation of four HIV antigen tests. J Virol Methods 2001;93(1–2):189–93.