Rapid ‘combination’ test fails to detect many cases of acute HIV infection

A rapid, point-of-care test, which aims to reduce the ‘window period’ through the detection of both antibodies and p24 antigen has poor performance in a clinical setting, researchers report in an article published online ahead of print by the Journal of Infectious Diseases. Whilst the test reliably identified people with established HIV infection, it delivered false negative results to the majority of individuals who were very recently infected with HIV.

Furthermore, this experience from Malawi echoes the findings of British researchers, published last year in Sexually Transmitted Infections.

Rapid diagnostic tests, which can be operated at the point-of-care by doctors or nurses, offer numerous advantages over conventional laboratory tests. Such devices are essential in resource-limited settings where laboratories are inaccessible or unavailable. Furthermore, as patients do not need to come back for their results on another day, far more people receive their results.

However almost all rapid tests only detect HIV antibodies, which are only detectable approximately one month after infection. In contrast, tests conducted in a laboratory may also be able to detect p24 antigen or HIV RNA, both of which are detectable earlier than this. These tests therefore shorten the ‘window period’.

Introduced in 2009, the Determine HIV-1/2 Ag/Ab Combo is the only rapid test which combines detection of antibodies and p24 antigen.

Nora Rosenberg and colleagues wished to monitor the performance of the Determine Combo test at two clinical sites in Lilongwe, Malawi. They compared the test’s results with those of other, validated tests.

Of 1009 individuals who tested, 838 were HIV-negative, 163 had established HIV-infection and 8 had acute infection (in other words, they were infected in the past few weeks). Individuals were judged to have acute infection if a laboratory test showed that they had HIV RNA but had had at least one negative result for HIV antibodies.

Among those with established HIV-infection, the test performed well. All but one individual was correctly identified, giving a sensitivity of 99.4% (95% confidence interval 96.6% - 100%).

However among the eight people with acute infection, only two were identified by the test, giving a sensitivity of 25% (95% confidence interval 3.2% - 65.1%). Moreover, these two individuals tested positive for HIV antibodies, and not for p24 antigen. While this does suggest that the antibody part of the test performs better than some other antibody tests, the p24 antigen part of the test performed poorly – but the detection of p24 antigen is key to identifying people who have recent infection.

Moreover, when the test was performed on 838 HIV-negative samples, 14 of the results for p24 antigen were false positives, further undermining the usefulness of the test in this setting.

The authors say that in this field evaluation “the sensitivity and specificity of the p24 antigen component of the test were inadequate for widespread use for detecting acute HIV infection.”

It is possible that a point-of-care test in the ‘lateral flow’ format has a much lower level of sensitivity than modern laboratory tests, which are able to detect lower concentrations of p24 antigen. An alternate possible explanation is that the test performs poorly with clade C HIV-1, the predominant subtype in Malawi.

However disappointing results have also been reported by Julie Fox of Guys’ and St Thomas’ Hospitals, London (where clade C is less common). She tested the Determine Combo test on stored samples from 36 people with HIV, each of whom had detectable p24 according to laboratory tests. Ten of them had detectable antibodies, while 26 did not (in other words, they had acute infection).

Overall, the rapid test detected p24 antigen for only half those tested, giving a sensitivity of 50% (95% confidence interval 34% - 66%).

The test failed to detect HIV infection (either via p24 antigen or antibodies) in ten of 36 cases, nine of whom had acute infection.

The researchers note that there are significant public health consequences to giving a negative HIV test result to a person whose infection is at its most infectious stage.

Specifically, Fox and colleagues recommend that combination laboratory tests are used for individuals with suspected acute infection and that health providers explain the limitations of tests to patients.