PCR detection of viral RNA (and/or proviral DNA in PBMCs) can be ultra-sensisitve and able to detect infection when the virus is being so well supressed by antiretroviral therapy that the infected person is not creating a detectable high-titer antibody response to the infection.   However, PCR is also quite prone to false-positive results, especially when it is being pressed to the limit to detect a very few infected cells.

In any case like this, which is very rare and not a routine matter, individuals need to weigh many factors in making choices.  The downsides of continuing therapy are cost, potential negative side effects, inconvenience and so on.  The downside of discontinuing therapy is that if there is virus present it can replicate in the absence of treatment and the individual will now be similar to any other infected individual, requiring re-starting therapy and so on.

So the first issue is how much effort should or can be put into proving for sure that the PCR positive results are genuine and not false-positive.  Using double-blinding of samples and controls from known negatives and from a known positive diluted to be just above the theoretical detection limit could be a relatively simple way to make sure the patient is truly PCR-positive.  Several replicates of patient sample and each of the positive and negative controls should be created and double-blinded so that the lab doing the PCR has no knowledge of which samples are which. 

Another alternative would be to sequence the PCR product from the infant and check to make sure that it matches the mother.  If the PCR is false-positive it is far more likely to be from random contaminating DNA than from contamination with the mother's virus unless the mothers samples have been amplfied in this lab along with the infant's positive result samples.

There should not be any one "gold standard" for all determinations.  There will always be some special cases where it can be worth spending thousands of dollars to be "certain" of a positive or negative diagnosis, while for the vast majority of average cases a determination can be made with greater than 99.9% certainty for less than $100.  It can be very difficult to make cost:benefit decisions when a human life is at stake, or even an issue of human health decision.  Decisions for the general public from a government department of public health point of view are often contradictory to what is best for any one individual.  Decisions for any one individual depend upon many factor about that individual such as their access to treatment, their sensitivity or tollerance of drug therapy, their finances, their age, etc.

PCR detection of viral RNA (and/or proviral DNA in PBMCs) can be ultra-sensisitve and able to detect infection when the virus is being so well supressed by antiretroviral therapy that the infected person is not creating a detectable high-titer antibody response to the infection.   However, PCR is also quite prone to false-positive results, especially when it is being pressed to the limit to detect a very few infected cells.

In any case like this, which is very rare and not a routine matter, individuals need to weigh many factors in making choices.  The downsides of continuing therapy are cost, potential negative side effects, inconvenience and so on.  The downside of discontinuing therapy is that if there is virus present it can replicate in the absence of treatment and the individual will now be similar to any other infected individual, requiring re-starting therapy and so on.

So the first issue is how much effort should or can be put into proving for sure that the PCR positive results are genuine and not false-positive.  Using double-blinding of samples and controls from known negatives and from a known positive diluted to be just above the theoretical detection limit could be a relatively simple way to make sure the patient is truly PCR-positive.  Several replicates of patient sample and each of the positive and negative controls should be created and double-blinded so that the lab doing the PCR has no knowledge of which samples are which. 

Another alternative would be to sequence the PCR product from the infant and check to make sure that it matches the mother.  If the PCR is false-positive it is far more likely to be from random contaminating DNA than from contamination with the mother's virus unless the mothers samples have been amplfied in this lab along with the infant's positive result samples.

There should not be any one "gold standard" for all determinations.  There will always be some special cases where it can be worth spending thousands of dollars to be "certain" of a positive or negative diagnosis, while for the vast majority of average cases a determination can be made with greater than 99.9% certainty for less than $100.  It can be very difficult to make cost:benefit decisions when a human life is at stake, or even an issue of human health decision.  Decisions for the general public from a government department of public health point of view are often contradictory to what is best for any one individual.  Decisions for any one individual depend upon many factor about that individual such as their access to treatment, their sensitivity or tollerance of drug therapy, their finances, their age, etc.

We should appreciate Dr. Brian Thomas for his excellent explanation and reply for the query raised by Dr. M.K.Saha,

This is the condition, where the babies immune system is being already suppressed to produce detectable antibodies in circulating blood system.

Since the ART treatment is going on, by nature the viral replication and infection of other cells is being depleted, so the antibody detection would leads to false negative.

Viral can be present in any part of the babies cells and can rejunavate even at later stage.

Hence, Continues RT-PCR for every one month sample will tell us the viral load in the babies blood.

Which will be the right way i suppose.

Sorry, if any wrong statement i made.

thank you.

This is a rare situation. Need to think rarest rare possibilities. One such can be Primary Immunodeficiency ( B cell deficiency). A simple test a undertake a serum globulin or  revere blood group ,  if . baby is not 'AB' blood Group.

This phenomenon Iimmunosilent) is also reported with HCV. (Nucleic Acid Positive, Antibody negative).

Dear Colleague,

I wish to respond first to your last question.

1. In infancy and up to 18 months of age, HIV DNA PCR is the confirmatory test for HIV infection; antibody tests in this age range can only serve as screening tests for resource constrained settings. This is especially important where a baby is being breastfed by an HIV infected mother because of on-going transfer of both virus and maternal antibodies through breast milk. After 18 months antibody tests are reliable for diagnosis if the quality of the testing methodology or kits are guaranteed. DNA PCR however remains the gold standard to resort to in the event of antibody tests becoming unreliable - due to issues such as questionable test quality (e.g. kit expiry, poor test kit storage conditions, poor/incorrect technique, etc). 

2. There are a number of possibilities to explain why an infant that tested positive by PCR at 6 months, could have a negative HIV later at 18 months. But before considering these possibilities it is very paramount to consider the clinical status (disease staging) of both the child and mother because clinical features/diagnostic parameters can be informative for a holistic approach:

a. Could it be due to a problem with the HIV antibody test reagents or kits giving false positive results?

b. Could it be a technical error on the part of the lab staff who performed the test?

c. Could it be due to errors in labelling specimens from the point of collection where labels/form were inadvertently swapped for different patients?

d. Could it be an error in recording test results after test were completed, perhaps yet again by inadvertent swap of results? 

A DNA PCR has higher specificity than antibody tests in general, and so it is more likely for some of the errors to arise from problems with the antibody testing.

It will be wise in this case to do antibody re-test with another set of reagents or kits (a different batch entirely, or even a different product by another manufacturer) and if results are negative, consider a 3rd test kit for a "tie breaker" (as recommended in serial HIV testing strategy). A consideration could also be made for ultimately repeating the DNA PCR as well if there is still some uncertainty.

A positive DNA PCR confirms infection especially if the inadvertent errors outlined above are ruled out.

"Sero-conversion" where a baby confirmed to be DNA PCR positive became PCR negative is very rare especially recalling that only very few case reports were documented in the literature (mostly in the 90s) and most recently, in the case of the mysterious "Mississippi baby". 

Regarding whether to continue ARV therapy or not, it is pertinent to immediately determine the correct HIV status of this child; one could maintain child on ART for the interim period of the laboratory evaluation (before results become available).  

Immunostaining (ELISA) assay is normally negative for infants whose HIV-positive mothers were under HAART during baby's prenatal period, there are hundreds of such cases reported from South Africa. All kits for ELISA HIV tests contain short HIV recombinant peptides from most conservative virus genes such as gag,pol, nef, not antibodies to HIV. Antibodies are detectected in human blood serum, and of course this immune response for HIV is detectable only after viral load in bloodstream rises to hundreds of copies per serum ml - for the children from HAART-treated mothers this normally happens after 18 months old and in some cases after 3 y.o. RT-PCR or RealTime RT-PCR are being used because these methods can measure viral load (HIV RNA copies number) quantitatively. RT-PCR positive results should appear at least several weeks and preferably several months before infant from HIV-positive perants becomes ELISA sero-positive.

DNA-PCR diagnostic is the only trustful method to distinguish HIV presence in general, especially for high-risk children. By now there are many cases babies from HAART-treated mothers stay uninfected completely, so DNA HIV diagnostic is worth trying, especially because it is not necessary to repeat it more than two times for one person if next infection catching chance is not under the investigation. The next level of guarantee of absence of mistake is DNA-PCR with following cloning and HIV sequence, but it is reasonable only if PCR is positive.

To start child's ART treatment before viral load (RT-PCR data) is stable or grows over 500-1000 copies/ml is nonsence. All types and combinations of HAART have multiple severe bystander effects which are doubled and tripled in children.

Dr. Foley has surprisingly vogue idea in the field, and very little practical knowledge.

Dear Dr Lawal, 

I strongly agree with your comments and really felt so useful information you have shared.

thanks a lot Doctor.

All theory has been spread out on this case, but I think you want practical advice. You have two  conflicting test results: one positive  HIV DNA  PCR  at 6 m age  and  one  negative antibody result at 18 months. What you  have to do  is to repeat the PCR.  Chances  are high that it will  now also be  negative. In that case, the most likely interpretation is that the first PCR result  was  false-positive;  consequently stop  ART. Repeat testing (antibody test  suffices) three  months later.  If negative,  it is  proven that the baby was and is uninfected and  that the  PCR was  false  positive.