Departments of Internal Medicine,
1 Medical Microbiology, Onze Lieve Vrouwe Gasthuis (OLVG), 1091 HA Amsterdam, The Netherlands
2
Received 21 October 2002/ Returned for modification 30 December 2002/ Accepted 15 May 2003
We evaluated the reliability of a rapid human immunodeficiency
virus type 1 test for quick clinical decision making, such as
in needle-stick accidents. The test was evaluated with 1,160
patients. It proved to be a simple and useful test with 99.6%
specificity and 99.4% sensitivity. One patient with late-stage
AIDS had a false-negative result.
In needle-stick accidents and clinical situations with a need
for immediate human immunodeficiency virus (HIV) testing, rapid
tests to detect HIV antibodies facilitate therapeutic and preventive
strategies. The Onze Lieve Vrouwe Gasthuis (OLVG) hospital is
an HIV referral center situated in the city center of Amsterdam,
which has a high HIV seroprevalence (
8). In this setting, there
is a need for a rapid HIV test. The Determine HIV 1/2 test (Abbott
Diagnostic Division, Hoofddorp, The Netherlands) is a rapid
immunochromatographic test for the detection of antibodies to
HIV types 1 and 2 (HIV-1 and -2, respectively). The performance
of this test has been described previously for patients from
Thailand and Africa (
1,
2,
5). We examined this test in a Dutch
population. Between July 1999 and May 2001, 1,160 consecutive
patients were tested for HIV antibodies. Fresh serum samples
were examined by both rapid test (RT) and a microparticle enzyme
immunoassay (MEIA). In case of positivity of one or both tests,
a line immunoblot assay was performed. If RT and MEIA results
were disconcordant, the tests were repeated immediately. Additionally,
these patients were evaluated for risk factors for HIV infection
and followed up serologically and/or clinically, if necessary.
In a few cases, HIV load was measured in EDTA plasma with the
HIV-1 RNA 3.0 assay (bDNA; Bayer Diagnostics, Fernwald, Germany)
with a detection level of 50 copies of RNA per ml. The following
HIV antibody tests were used. (i) The Determine HIV 1/2 test
is a rapid immunochromatographic test that uses either serum,
plasma, or whole blood. The results can be read after 15 min,
and no special equipment is needed. (ii) The AXSYM HIV 1/2 gO
(Abbott) is a microparticle enzyme immunoassay. The cutoff rate
is calculated from the mean rate of three index calibrator replicates.
A sample rate divided by a cutoff rate (s/co ratio) 1.00
is considered positive. (iii) The Inno-Lia HIV test (Innogenetics,
Ghent, Belgium) is a line immunoblot assay.
All assays were carried out as described by the manufacturers.
For 1,149 of the 1,160 patients (99%), RT and MEIA showed concordant results (988 negative and 161 positive results). In 11 cases, RT and MEIA results were disconcordant (1%; Table 1). We consider 10 of the 11 disconcordant cases to be false positives, with 6 false-positive MEIA results (patients D, F, G, H, J, and K) and 4 false-positive RT results (patients B, C, E, and I). In six patients (B to G), a serological follow-up indicated no infection with HIV. In two patients (G and H), an HIV RNA level of <50 copies per ml ruled out infection with HIV-1. Theoretically, a beginning HIV-1 group O or HIV-2 infection cannot be excluded in patient H. However, this is very unlikely in a Dutch population. The other three patients (I, J, and K) had no risk factors for HIV infection and no signs or symptoms of the disease. All six false-positive MEIA samples had an s/co ratio just above the cutoff. Two patients with a false-positive MEIA had concomitant infections (an acute cytomegalovirus infection and an amebic liver abscess) that might be the cause of the false-positive results. False-positive enzyme-linked immunosorbent assays have been described before (3, 4, 6, 7), and cross-reactivity is one of the known causes. A false-negative MEIA did not occur. The reasons for false-positive RT reactions are unknown. The patients remained positive on repeated testing (Table 1).
fig.ommitted |
TABLE 1. Results of primary tests and follow-up of patients with discordant results in the rapid Determine HIV 1/2 test and the AXSYM MEIAa
| |
One patient (A) had a false-negative RT result. He was a 23-year-old
man, admitted with
Pneumocystis carinii pneumonia, oral thrush,
and general malaise. The HIV RNA level was 890,000 copies per
ml, the CD4 count was 0.02
x 10
-9/liter, and the HIV p24 antigen
concentration was 160 pg/ml. The MEIA was reactive, but the
RT was negative. The line immunoblot assay was indeterminate.
The absence of antibodies to antigens encoded by
gag and to
envelope protein gp120 was notable. Table
2 shows the serological
follow-up of this patient. Based on the follow-up, we assumed
that the positive tests in all disconcordant samples were false
positives, except for patient A. Thus, the MEIA had 100% sensitivity
and 99.4% specificity and the rapid Determine HIV 1/2 test had
99.4% sensitivity and 99.6% specificity.
fig.ommitted |
TABLE 2. Results of the serological follow-up of the patient with HIV infection and a negative rapid Determine HIV 1/2 test
| |
Previous studies of non-Western populations have shown a similar
specificity, but in contrast to our study, the sensitivity was
100% (
1,
2,
5). Aidoo et al. (
1) compared the Determine HIV
1/2 assay and Diagnostic HIV SPOT with a particle agglutination
test and a confirmational Western blot as the "gold standard"
in 200 specimens. The results revealed 125 HIV-positive patients
and 75 HIV-negative patients, with a specificity of 100% for
both tests and sensitivities of 100% for the HIV 1/2 assay and
98% for the HIV SPOT assay. Arai et al. (
2) had 398 samples
from Thailand and 100 samples from Ivory Coast from patients
proven to be HIV positive by EIA and confirmed by Western blotting.
The Determine HIV 1/2 test showed 100% sensitivity and specificity.
Koblavi-Deme et al. (
5) compared four rapid assays of 1,216
sera with known HIV serological status. All four rapid tests
showed 100% sensitivity and specificities ranging from 99.4
to 100%: Determine HIV 1/2, 99.4%; HIV SPOT, 99.6%, Capillus,
99.7%; and Genie II, 100%. The patient with the false-negative
RT result turned out to have end-stage AIDS. The RT apparently
lacks sufficient antigenic determinants to detect all patients
with end-stage HIV infection.
In summary, the HIV 1/2 rapid test has proven to be a rapid, simple, and useful test for the detection of HIV antibodies. Because of the ability to test immediately, the use of this test with needle-stick accidents has led to an impressive reduction in prescription of postexposure prophylaxis and sick leave in our hospital (8). However, the clinical situation should be taken into account when interpreting the test result because of the possibility of false-negative results in late-stage AIDS.
- Aidoo, S., W. K. Ampofo, J. A. M. Brandful, S. V. Nuvor, J. K. Ansah, N. Nii-Trebi, J. S. Barnor, F. Apeagyei, T. Sata, D. Ofori-Adjei, and K. Ishikawa. 2001. Suitability of a rapid immunochromatographic test for detection of antibodies to human immunodeficiency virus in Ghana, West Africa. J. Clin. Microbiol. 39:2572-2575.
- Arai, H., B. Petchclai, K. Khupulsup, T. Kurimura, and K. Takeda. 1999. Evaluation of a rapid immunochromatographic test for detection of antibodies to human immunodeficiency virus. J. Clin. Microbiol. 37:367-370.
- Celum, C. L., R. W. Coombs, M. Jones, V. Murphy, L. Fisher, C. Grant, L. Corey, T. Inui, M. H. Wener, and K. K. Holmes. 1994. Risk factors for repeatedly reactive HIV 1 EIA and indeterminate Western blots. Arch. Intern. Med. 154:1129-1137.
- Cordes, R. J., and M. E. Ryan. 1995. Pitfalls in HIV testing. Application and limitations of current tests. Postgrad. Med. 98:177-180, 185-186, 189.
- Koblavi-Dème, S., C. Maurice, D. Yavo, T. S. Sibailly, K. N'Guessan, Y. Kamelan-Tano, S. Z. Wiktor, T. H. Roels, T. Chorba, and J. N. Nkengasong. 2001. Sensitivity and specificity of human immunodeficiency virus rapid serologic assays and testing algorithms in an antenatal clinic in Abidjan, Ivory Coast. J. Clin. Microbiol. 39:1808-1812.
- Mylonakis, E., M. Paliou, M. Lally, T. P. Flanigan, and J. D. Rich. 2000. Laboratory testing for infection with the human immunodeficiency virus: established and novel approaches. Am. J. Med. 109:568-576.
- Profitt, M. R., and B. Yen-Lieberman. 1993. Laboratory diagnosis of human immunodeficiency virus infection. Infect. Dis. Clin. N. Am. 7:203-219.
- Regez, R. M., P. J. G. M. Rietra, C. T. M. van der Linden, P. H. J. Frissen, H. M. Weigel, and K. Brinkman. 2002. Reducing the risk of blood-transmitted infections of HIV and hepatitis B or C virus in a teaching hospital in Amsterdam: evaluation of a protocol for needle accidents among hospital staff during the period 1997-2001. Ned. Tijdschr. Geneeskd. 46:617-621.
作者:
G. E. L. van den Berk P. H. J. Frissen R. M. Reg 2007-5-10