False-negative confirmatory HIV tests in children are frequent after starting ART

False-negative rapid antibody tests, in the absence of virologic testing at the age of 18 months, in children receiving antiretroviral treatment can result in interruption of treatment, and false-negative results occur with high frequency, according to the results of a study carried out in Lesotho, published in the advance online edition of AIDS.

Sixty per cent of children with negative or discordant post-18 month confirmatory rapid HIV tests were definitively positive by HIV-DNA PCR while only 4% were definitively negative, Garcia-Prats and colleagues report in this retrospective review of routine programme data from an HIV clinic in Maseru, Lesotho.

Starting ART at an early age (under nine months) was significantly associated with a false-negative or discordant rapid test result (OR=4.25, p=0.002). 94% (46/49) of these children with a non-positive result were on ART.

In examining the challenges of post-18 month confirmatory HIV testing the authors highlight this common occurrence of false-negative rapid tests which can lead to the inappropriate interruption of treatment in children with HIV, potentially leading to disease progression and the development of resistance.

In settings of high HIV prevalence (over 5%) a two-test serial testing strategy is recommended for diagnosis of HIV.

In resource-poor settings, diagnosis of HIV in adults and children over 18 months of age is based on two positive HIV antibody tests and includes rapid antibody tests (RT) or enzyme immunoassay (ELISA) antibody tests.

Among young infants, the risks of HIV transmission through breastfeeding, together with the transfer during pregnancy of maternal antibodies that can persist until 18 months of age further complicates HIV testing.

Virologic testing with HIV DNA polymerase chain reaction (PCR) is recommended to confirm HIV infection in this group.

In resource-poor settings, financial or other constraints restrict the use of PCR either for a first test or for a second confirmatory one. In the first instance, the World Health Organization (WHO) recommends clinical diagnosis to guide the use of ART in infants under 18 months of age. In both circumstances WHO 2006 guidelines recommends HIV antibody testing after 18 months of age to confirm the initial diagnosis. 

In spite of this recommendation, scant data exist assessing the outcomes of post-18 month confirmatory testing among infants starting ART immediately.

So the authors chose to look at the challenges of such testing among infants enrolled at the Baylor Children’s Clinical Center of Excellence (COE), a paediatric and family HIV clinic in Maseru, Lesotho.

Lesotho has a 23.6% adult HIV prevalence, the third highest worldwide, an estimated 28,000 children under the age of 15 living with HIV and 14,000 children are born to mothers who have HIV each year.

HIV testing at the clinic follows national guidelines. In April 2006, a national programme for early infant diagnosis using dried blood spot (DBS) HIV DNA PCR began. Guidelines allowed for a single positive PCR test to be done for children under 18 months of age and recommended antibody testing after 18 months of age for confirmation.

Findings are based on three illustrative case studies and a retrospective chart review.  Children included those enrolled from December 2005 until January 2009 with a documented positive HIV DNA PCR at under 18 months of age and documented HIV rapid tests (RTs) after 18 months of age.

Definition of a definitive HIV status was determined as follows:

• Definitively HIV positive – two positive PCRs, or one positive PCR with either a positive post-18 month ELISA antibody test or two positive post-18 month rapid antibody tests.
• Children were considered to be probably positive with one positive PCR and discordant post-18 month rapid antibody tests.
• Children were considered definitively negative if they did not meet the above criteria and after test results and clinical details were reviewed.
• Children not meeting any of the above criteria were considered undetermined.

Among the 109 children included 49 (45%) had negative (22) or discordant (27) confirmatory rapid tests.

Twenty of the 22 children with negative confirmatory rapid antibody tests were on ART. The mean age at the start of ART was 8.0 months, standard deviation 3.6, range 2.4-15.7 months. Of these 22, 60% (9/15) with follow-up PCRs were negative. Of these nine, four were found to be positive after ELISA HIV antibody testing.

Among the 49, 29 (60%) were definitively positive, 17 (35%) probably positive while only two were definitively negative and one of undetermined status.

While young age at the start of ART was significantly associated with false-negative results the authors note this is not just a problem among the youngest infants.

The authors acknowledge that incorrect testing procedures – rapid antibody tests in field settings done by non-laboratory personnel – may have contributed to their findings but cannot explain the high rates of false-negatives.

Their findings, they add, contribute to the phenomenon of seroreversion in children on effective ART but within the context of confirmatory testing in resource-poor settings.

Seroreversion, it has been hypothesized, among young children starting ART is due to the suppression of viral replication during a time when immune responses are impaired so decreasing antigenic stimulation.

Previous reports, note the authors, show that rapid antibody tests may be less sensitive than ELISA at lower antibody levels, so explaining positive ELISA after negative rapid antibody tests. However, loss of detectable antibody also happens in children on effective ART for both HIV ELISA and Western blot testing, they add.

The authors note that ART should not affect DNA PCR results, yet their findings showed children identified as HIV-infected through HIV DNA PCR having subsequent negative PCRs after commencing antiretroviral treatment.

Studies, they add, have shown detectable proviral HIV DNA in spite of prolonged viral suppression (under 50 copies/ml), yet in most of these a small percentage have proviral DNA levels below the limits of detection. They also cite the example of a child with known perinatal HIV infection who started ART at ten months of age and after three years of suppressive ART had negative HIV ELISA, Western blot and PCR testing. 

“Low copy number of proviral DNA is the most frequent explanation for false-negative PCRs,” they state. They suggest that dried blood spot sample collection may influence the detectability of HIV DNA in cases where HIV DNA copy numbers are low due to early initiation of antiretroviral therapy.

Limitations of the study include a lack of standardisation of post-18-month testing. Viral load testing is not routinely available so it was not possible to show definitively that false-negatives happened in children on ART with suppressed viral load.

The authors add that in line with WHO 2010 recommendations their findings support two virologic tests, where available, to establish a definitive diagnosis before a lengthy time on ART, so avoiding post-18-month confirmatory testing.

In children over 18 months of age on ART and without a confirmed diagnosis, the authors suggest that rapid antibody tests will provide a definitive diagnosis in some, but ELISA testing may be better and HIV DNA PCR may be of use in others.

However, even if all confirmatory tests are negative, stopping ART needs to be done with caution, they add, and the children closely followed up and repeat testing done to establish a definitive diagnosis.

The authors conclude “there is an urgent need for point of care HIV tests for use in children under 18 months of age so increasing access to early infant diagnosis, limiting the need for a presumptive clinical diagnosis and make initial and confirmatory testing simpler and quicker…and ensure all children receive accurate HIV tests for such a life-changing diagnosis.”