Kabamba et al.  and Delforge et al.  raised concern about the sensitivity of fourth-generation assays for the detection of primary HIV infection. In particular, one assay showed a very poor performance in two cases of HIV seroconversion. In both reports, paired samples were tested with older combined antigen/antibody (fourth-generation) assays launched between 1997 and 2000. With the exception of one test kit, all these assays exhibit a relatively high detection limit of the HIV antigen module (> 30 pg p24 antigen/ml). HIV-1 subtypes and HIV-2 are detected with a different, usually lower, sensitivity than with single p24 antigen assays . Therefore, they cannot substitute single antigen determination with sensitive commercially available enzyme immunoassays. As approximately one-third of the solid phase is coated with monoclonal antibodies for HIV p24 antigen detection, there is a more limited binding capacity available for antibody detection in comparison with antibody only assays (third-generation). In consequence, fourth-generation assays yield a potential risk of a second diagnostic window in the transition phase between the declining p24 antigen level and the beginning of seroconversion .
Meanwhile, new fourth-generation HIV screening assays with improved sensitivity have been extensively evaluated . The detection threshold of 3–10 pg of p24 antigen/ml is within the range of single HIV p24 antigen assays. Their sensitivity for HIV p24 antigen in seroconversion panels is equal to that of single antigen assays. HIV-1 non-B subtypes are detected with a comparable sensitivity in dilutions of virus lysates as with p24 antigen assays.
Whereas results from case reports give only a partial view of the performance of an assay, statistical analysis of published data provides more objective results on the sensitivity of fourth-generation assays. Table 1 gives an overview of the results obtained so far with fourth-generation HIV assays in comparison with antibody assays in seroconversion panels [3–10]. The criteria for inclusion of published evaluations in our analysis were as follows: citation in Medline, testing of at least 10 seroconversion panels per study, the availability of individual data for each test and panel in terms of the number of the day of the first positive result. The calculation model for time delays between assays established by the Paul Ehrlich Institute  was used. The time delay between blood sampling points in commercially available seroconversion panels is on average relatively short (2–7 days) but may last up to more than 2 weeks for certain panels. This method considers that seroconversion is theoretically possible the following day after the last negative follow-up sample. The total number and the average of days of time delay for panels were calculated in comparison with the most sensitive assay. The statistical significance of the reduction for each test was determined using the Wilcoxon test for matched pairs.
New assays with a sensitive antigen detection module, such as VIDAS HIV DUO Ultra and Cobas Core HIV Combi showed a statistically significant reduction of the diagnostic window in comparison with the most sensitive third-generation assays (Prism Anti-HIV-1/2 and Genscreen HIV-1/2 version 2). VIDAS HIV DUO with a fivefold lower sensitivity for HIV antigen than VIDAS HIV DUO Ultra detected primary infection significantly earlier in comparison to third-generation tests, with the exception of one evaluation. For the less sensitive assays, especially for Vironostika Uniform Ag/Ab, which showed a poor performance in the above-mentioned case reports, the difference in sensitivity in comparison with a sensitive third-generation test was not significant.
The difference in performance is also influenced by the choice of seroconversion panels, especially if a relatively small number is ts tested. The time delay between antigen and antibody detection influences the final result considerably. For example, if a high number of panels with concomitant appearance of antigen and antibody in the first follow-up sample are tested, the difference in performance between fouth and third-generation assays may be underestimated. This potential bias is probably eliminated by the analysis of a relatively large database, as shown in Table 1.
The author is very grateful to Drs Melchior and Kleider, Roche Diagnostics, Penzberg, Germany, for statistical analysis.
1.Kalamba BM, Vandercam B, Bodeus M, Moreau M, Goubau P. An HIV seroconversion case: unequal performance of combined antigen/antibodies assays.AIDS
2.Delforge M, Dollé N, Hermans P, Liesnard C. Delayed primary infection diagnosis with a fourth-generation HIV assay.AIDS
3.Gürtler L, Mühlbacher A, Michl U, Hofmann H, Paggi GG, Bossi V, et al.Reduction of the diagnostic window with a new combined p24 antigen and human immunodeficiency virus antibody screening assay.J Virol Meth
4.Weber B, Gurtler L, Thorstensson R, Michl U, Muhlbacher A, Burgisser P, et al.Multicenter evaluation of a new automated fourth-generation human immunodeficiency virus screening assay with a sensitive antigen detection module and high specificity.J Clin Microbiol
5.Brust S, Duttmann H, Feldner J, Gürtler L, Thorstensson R, Simon F. Shortening of the diagnostic window with the new combined HIV p24 antigen and anti-HIV-1/2/0 screening test Enzygnost HIV Integral.J Virol Meth
6.Couroucé AM,and groupe de travail Rétrovirus de la S.F.T.S. Tests de dépistage combiné des anticorps anti-VIH et de l'antigène p24.Gaz Transfusion Sang
7.Ly TD, Laperche S, Couroucé AM. Early detection of human immonodeficiency virus infection using third- and fourth- generation screening assays.Eur J Clin Microbiol Infect Dis
8.van Binsbergen J, Siebelink A, Jacobs A, Keur W, Bruynis F, van de Graaf M, et al.Improved performance in seroconversion with a 4th generation HIV antigen/antibody assay.J Virol Meth
9.Weber B, Fall EMB, Berger A, Doerr HW. Reduction of Diagnostic Window by New Fourth-Generation Human Immonodeficiency Virus Screening Assays.J Clin Microbiol
10.Weber B, Gurtler L, Thorstensson R, Michl U, Muhlbacher A, Burgisser P, et al.Multicenter evaluation of a new automated fourth-generation human immunodeficiency virus screening assay with a sensitive antigen detection module and high specificity.J Clin Microbiol