Contribution of Combined Detection Assays of p24 Antigen and Anti-Human Immunodeficiency Virus (HIV) Antibodies in Diagnosis of Primary HIV Infection by Routine Testing

In reference to the paper of Weber et al. (1), we report here the contribution of a fourth-generation human immunodeficiency virus (HIV) diagnosis assay, the VIDAS HIV DUO (bioMérieux), to diagnosis of primary HIV infection earlier than the third generation.

During a recent 9-month survey, from April to December 1998, a study was conducted in five French public and private laboratories that had reported diagnosis of primary HIV infections. In France, the use of two different HIV screening assays is mandatory for testing for non-blood-donation HIV. In four of the five laboratories, part of the routine testing was done for an anonymous high-risk population. A total of 29,657 consecutive and unselected samples were tested for the presence of anti-HIV antibodies systematically with the VIDAS HIV DUO plus one of the following third-generation detection assays: Ortho Capture HIV-1/HIV-2 (Ortho Diagnostic Systems), Genscreen HIV1/2 (Sanofi Diagnostics Pasteur), Axsym HIV-1/HIV-2 (Abbott), Vironostika HIV Uniform II plus 0 v 3.3 (Organon Teknika), or COBAS Core anti-HIV-1/HIV-2 EIA DAGS (Roche Diagnostics).

Of the 29,657 samples tested, 29,002 were found to be negative by both the VIDAS HIV DUO and a third-generation assay. A total of 453 samples were reactive with both assays. Of these 453 samples, 449 were confirmed to be positive by Western blotting (New Lav Blot I and II; Sanofi Diagnostics Pasteur) and four were found to be negative by the same technology. Follow-up samples were reactive with both the VIDAS HIV DUO and Western blotting, suggesting that the patients from whom these samples came were infected at the time the initial samples were collected. The positive prevalence rate of this tested population was 1.52%.

With the VIDAS HIV DUO, 160 samples were found to be repeatedly reactive only by this assay. Both an HIV antigen assay and Western blotting were performed on these 160 samples. All 160 samples were shown to be negative by Western blotting. Seventeen of the 160 samples were confirmed to show primary HIV infection with a positive p24 antigen test. In addition, the HIV type 1 (HIV-1) RNA viral load was determined for 11 of the 17 samples and the samples were also found to be positive. Finally, HIV infection was confirmed for all 17 patients by determining complete Western blot profiles on follow-up samples. These additional tests allowed us to confirm that the 17 patients who were negative, by both a third-generation assay and Western blotting, were positive with the VIDAS HIV DUO due to a primary HIV infection.

With the third-generation tests, there were 42 samples that were found to be repeatedly reactive only with these assays. Western blotting was performed for these 42 samples, and none was confirmed to be positive when follow-up samples were used.

For comparison of the sensitivity and specificity of the VIDAS HIV DUO with those of the third-generation tests, samples were considered to be either HIV positive or negative according to the following decision trees. Samples were considered positive if (i) at least one screening assay was positively reactive, and it was confirmed by Western blotting; (ii) if the Western blot was negative, then a p24 antigen test was performed (only in the case of positive samples for the VIDAS HIV DUO); (iii) if either the p24 antigen and/or RNA assay was positive, then Western blotting was performed on a follow-up sample; (iv) in all cases, there was a final positive result by Western blotting. Samples were considered negative (i) if both screening assays were negative or (ii) a negative confirmation result followed an enzyme immunoassay-positive reaction using Western blotting and/or a p24 antigen test.

When both the sensitivity and specificity of the VIDAS HIV DUO were compared with those of the third-generation tests, there were some differences to be noted. According to the protocol of the study (referred above), the sensitivity of the VIDAS HIV DUO was shown to be 100%, which was statistically significantly higher than the 96.4% of the third-generation tests (McNemar test P value < 0.01). Therefore, the use of this assay enabled the diagnosis of infection in 3.6% additional HIV-infected patients. These results agree with the data of Weber et al. (1), demonstrating earlier diagnosis of HIV infection by fourth-generation assays. In contrast, the specificity of the third-generation tests were more specific at 99.86% and statistically significant when compared with the 99.51% specificity of the VIDAS HIV DUO (McNemar test P value < 0.01).

We gathered the clinical data of these 17 patients to draw a picture corresponding to various cases of primary HIV infection (Table 1). One of the patients showed no symptoms because he was tested for HIV in the context of a presurgical procedure. In this case, the VIDAS HIV DUO detected a completely unexpected primary HIV infection that would have been missed with the third-generation tests. Primary HIV infection was suspected in 8 of these 17 patients; therefore, a p24 antigen detection assay was originally prescribed. The remaining nine patients were later diagnosed as HIV infected by using a p24 antigen test due to the repeatedly positive results of the VIDAS HIV DUO test.

TABLE 1.

Biological and clinical data of 17 patients with primary HIV-1 infection

Patient no.	OD cutoffa		Delay in 3rd gen assay results (days)a	p24 antigen concn (pg/ml)	HIV-1 RNA (103 copies/ml)	Delay in Western blot results (days)h	Clinical symptoms	Risk factors
	VIDAS HIV DUO	3rd Genb assays
1	7.6	0.8	5	>200	500c	16	Fatigue, polyadenopathy	Heterosexual
2	6.6	0.2	4	>200	>2,000c	14	Fever, skin rash, thrombopenia	Heterosexual
3	27.4	0.9	7	>200	1.1d	7	Fever, headache, weight loss	IV drug user
4	4.4	0.5	3	80	>500d	11	Fever, headache	Homosexual
5	29.6	0.3	8	69	NTe	14	Skin rash, fatigue	Homosexual
6	5.4	0.3	7	23	NT	15	Fever, muscular fatigue, polyadenopathy	Heterosexual
7	35.1	0.4	9	>200	2,000f	16	Fever, skin rash	Homosexual
8	9.0	0.3	7	>200	NT	13	No symptoms	Heterosexual
9	4.8	0.2	6	168	NT	16	Fever	Heterosexual
10	14.5	0.5	4	>200	NT	10	Fever, skin rash	Homosexual
11	21.6	0.2	5	>200	27,500c	10	Fever, hepatomegalia	IV drug user
12	31.4	0.3	7	>200	NT	12	Fever, polyadenopathy, hepatomegalia	Heterosexual
13	41.4	0.5	2	>200	211c	30	Skin rash, fever, oral candidose	Heterosexual
14	3.8	0.8	4	>200	4,800c	15	Fever, infectious mononucleosis symptoms	Heterosexual
15	16.5	0.3	3	>200	3,500c	15	Fever, vulvo vaginitis	Heterosexual
16	16.4	0.5	21	>200	394c	21	Fever, weight loss	Heterosexual
17	12.4	0.3	5	>200	3,000c	11	Fever, muscular fatigue	Homosexual

^aRatio > 1, positive; ratio < 1, negative. 

^b3rd Gen, third generation. 

^cDetermined by AMPLICOR HIV assay (Roche). 

^dDetermined by bDNA HIV assay (Chiron version 2.0). 

^eNT, not tested. 

^fDetermined by NucliSens HIV assay (Organon). 

^gMean delay in results of third-generation assays relative to those of the VIDAS HIV DUO, 6.29 days. 

^hMean delay in results of Western blotting relative to those of the VIDAS HIV DUO, 14.47 days. 

In conclusion, these results clearly demonstrate that by using the most up-to-date diagnostic tools, such as these fourth-generation tests, we can reduce the risk of missing any unsuspected primary HIV infections in routine testing. Since the viral load is very high during primary HIV infection, we now have the capability of diagnosing the HIV infection earlier, allowing for the administration of antiretroviral therapy as early as possible. This earlier diagnosis gives us the possibility of increasing the efficiency of the treatment and especially helps to decrease the risk of HIV transmission.