Identification of recent human immunodeficiency virus (HIV) infection is an important tool for monitoring HIV transmission. During the last few years, several serological assays have been developed for this purpose and used in cross-sectional studies.1 The serological testing algorithm for recent HIV seroconversion (STARHS) was developed in 19982 and used with the Abbott HIVAB 3A11 assay (Abbott Laboratories, Abbott Park, Chicago, Illinois, USA) and with the Vironostika HIV-1 Microelisa System (bioMérieux SA, Marcy l’Etoile, France). The sensitivity of both assays was lowered in order to obtain a negative result in specimens with low antibody titers, such as those from individuals with a recent infection. These assays are no longer available, and laboratories have turned to new methods. The BED assay (Calypte Biomedical Corporation, Portland, Oregon, USA) measures anti-HIV IgG titers3 and includes a calibrator to ensure comparability of results. Furthermore, the BED assay is included in an external quality program offered by the Centers for Disease Control and Prevention (CDC, Atlanta, Georgia, USA).

In Catalonia, the STARHS was introduced using the Vironostika assay as part of the enhanced HIV/STI surveillance program in 2003. As the Vironostika assay has no longer been available since 2007, our laboratory changed to the BED assay in 2008. Our aim was to assess whether the results obtained by both techniques were comparable.

A total of 101 serum specimens from HIV-1–positive individuals were selected from a previous study.4 The selection criteria were sufficient sample volume and minimum available clinical and laboratory information (CD4+ T-cell count, HIV viral load, HIV infection stage, previous antiretroviral treatment). Patients were classified into three groups. The first group included 15 recently infected individuals (sera were drawn no more than 6 months after seroconversion in nine patients, and the rest had a diagnosis of acute HIV infection). The second group included three patients with long-standing infection (subjects infected for >12 months) and 45 patients with a diagnosis of AIDS (clinical criteria or CD4+ T-cell count under 200 cells/μl). Finally, 38 specimens were unable to be classified as either recent infections, long-standing infections, or AIDS.

The agreement between Vironostika and BED assays was good (κ=0.738, P<.005), which is consistent with the results of two published studies.5,6 Sensitivity to detect recent infection was 93.3% (95% CI: 68.1 – 99.8) for both the Vironostika and the BED assays. Specificity for detecting long-term infections was 75.0% (95% CI: 60.4 – 86.4) using the Vironostika, and 81.3% (95% CI: 67.4 – 91.1) using the BED. Positive predictive values were 60.9% (95% CI: 38.5 – 80.3) using the BED, and 53.8% (95% CI: 33.4 – 73.4) using the Vironostika. Negative predictive values were 97.5% (95% CI: 86.8 – 99.9) using the BED, and 97.3% (95% CI: 85.8 – 99.9) using the Vironostika. Table 1 shows the samples identified as recent infections by both BED and Vironostika according to their clinical characteristics and laboratory information.

The BED assay correctly classified a greater proportion of recent infections and patients with AIDS than Vironostika. These results are similar to those of a previous study,6 in which the Vironostika kit also tended to misclassify more individuals with long-standing infections or AIDS as recently infected in a comparison with the avidity index method.7 The misclassification of patients with AIDS or CD4+ T-cell counts ≤200 cells/μl is explained by the low anti-HIV IgG titers. Hence the importance of excluding those samples belonging to patients fulfilling these criteria from STARHS testing when this information is available.

The BED assay offers several advantages over the Vironostika: i) it has been reported to have better reproducibility, since it is based on the HIV IgG/non-HIV IgG ratio and uses a simple 1:100 dilution;8 ii) the BED assay can also be automated providing more precise results than the Vironostika assay, which is performed manually; and iii) the window period of the Vironostika assay differs for B and non-B HIV-1 subtypes, whereas these differences are less pronounced with the BED.9 For all those reasons, the BED assay offers a good alternative to the discontinued Vironostika assay.