A wide range of point-of-care tests
have been manufactured in many countries, but only a few of them have been
subject to rigorous, independent evaluations, and even fewer are marketed in
the UK. Research on HIV tests is only occasionally published in medical
journals. Informally, laboratory professionals may have insights into which
tests perform best.
It is important to verify that any test used is CE marked.
This should mean that the test conforms to European health and safety
legislation, although it does not necessarily mean that test performance has
been independently evaluated.
There
are variations in accuracy from one test to another, with some older
tests that are not usually marketed in the UK having a sub-optimal sensitivity
and specificity. However, evaluations by the World Health Organization
of several rapid diagnostic tests
that either have CE marks or are approved by the US Food and Drug Administration (FDA), indicate that most are
extremely accurate. The key measures of accuracy are sensitivity (the
percentage of results that are correctly positive when HIV is actually present)
and specificity (the percentage of results that are correctly
negative when HIV is not present).
Of note, in the World Health Organization data below, the tests
were performed with samples of plasma or serum. However, the tests are less sensitive
when testing whole blood sampled from a finger prick. Moreover, the blood was
taken from people who had chronic (not recent) HIV infection, but the tests are
less accurate in cases of recent infection.
Test
|
Detects
|
Sensitivity
|
Specificity
|
OraQuick
HIV-1/2 Rapid HIV-1/2 (OraSure)
|
IgG
|
98.1%
|
100%
|
HIV 1/2 STAT-PAK (Chembio)
|
IgG
|
99.5%
|
100%
|
Alere Determine HIV-1/2
(Alere)
|
IgG + IgM
|
100%
|
98.9%
|
Uni-Gold
HIV (Trinity)
|
IgG + IgM
|
99.8%
|
99.9%
|
INSTI HIV-1/HIV-2 Antibody Test
(bioLytical)
|
IgG + IgM
|
100%
|
99.7%
|
SD BIOLINE HIV-1/2 3.0
(Standard Diagnostics)
|
IgG + IgM
|
99.8%
|
99.8%
|
DPP® HIV 1/2 Assay (Chembio)
|
IgG
|
99.8%
|
99.8%
|
VIKIA HIV 1/2 (bioMérieux)
|
IgG
|
99.4%
|
99.9%
|
Reveal Rapid HIV Antibody Test (MedMira)
|
IgG
|
99.8%
|
99.9%
|
Whereas the data above relate to blood samples, the OraQuick
Advance Rapid HIV-1/2 is a widely used test which can also test oral
fluid samples. While this is non-invasive and highly convenient for the person
testing, performance
is slightly poorer when testing oral fluid samples than when testing blood
samples. The key reason that there are lower quantities of HIV
antibodies in oral fluid than in whole blood, especially after recent
infection. In seven studies which made a direct comparison of test performance,
the pooled sensitivity with oral fluid was 98.03%, compared to 99.68% with
whole blood.
There is one rapid, point-of-care test that looks for both
antibodies and p24 antigen, in a similar way to antibody/antigen laboratory
tests. The Alere Determine HIV-1/2 Ag/Ab
Combo was originally introduced in 2009, with an updated version called the
Alere HIV Combo launched in Europe in
2015 (the older version is still marketed in the United States and in some
parts of the world).
The promise of having a ‘fourth-generation’ point of care
test that detects p24 antigen is that the window period should be shortened. However, several
studies found that although the older version of this test performed well in
respect of established HIV infection, its ability to detect recent
HIV infection did not match that of laboratory antibody/antigen tests. The
test was quite insensitive to p24 antigen, making it only marginally better
than antibody-only tests in detecting acute (recent) infection.
The handful of studies published so far on the newer version
suggests it has better performance in acute infection, although it still does
not match that of antibody/antigen laboratory tests. The Alere HIV Combo’s sensitivity during acute infection has been
variously estimated to be 28% (in three African
countries), 54% (France),
65% (the Netherlands)
and 88% (UK).
An
analysis pooled the results of 18 separate studies in which a
point-of-care test (including Determine,
OraQuick, UniGold and INSTI) was
compared with a more sensitive laboratory test. Compared with fourth-generation
laboratory tests, the estimated sensitivity of the point-of-care tests was
94.5% (95% confidence interval 87.4-97.7) and specificity was 99.6% (99.4-99.7).
Compared with RNA (viral load) tests, the estimated sensitivity was 93.7% (95%
confidence interval 88.7-96.5) and specificity 98.1% (95% CI: 97.9-98.2).
Sensitivity was higher in nine studies conducted in African
countries than in the nine studies conducted in the United States and other
wealthy countries. This is likely to be due to different populations coming
forward for screening. Whereas 4.7% of those testing positive in African
studies had acute (recent) HIV infection, this figure rose to 13.6% in the
high-income countries.
A
study in five African countries found that the performance of point-of-care
tests was sub-optimal. Samples from some countries were more likely
to have false positive results than others, suggesting that tests need to be
locally validated and that some tests may be more accurate in relation to some
HIV subtypes than others. The researchers found a high number of false positive
results, whereas false negative results were relatively rare. The specificities
of the First Response HIV Card Test
1–2.0, INSTI HIV-1/HIV-2 Antibody Test, Determine HIV-1/2 and Genie Fast HIV 1/2 were all between 90
and 95%. The findings confirm that the diagnosis of HIV should not be based on
results from a single HIV rapid diagnostic test. A combination of HIV tests,
and more specifically an algorithm (sequence) of two or three different tests,
is required to make an HIV-positive diagnosis. This is recommended in testing
guidelines.
All HIV tests need to have reactive (preliminary positive)
results confirmed with confirmatory tests. A particular challenge healthcare
workers have with rapid tests is how to communicate a reactive result
to the person testing (who may be present while the result is being read) and
explain that supplementary tests are needed. These problems are less frequently
faced with laboratory testing – a large enough blood sample was taken to allow
for it to be tested several times and for uncertainties in the diagnosis to be
resolved.