There has been great excitement in the media and in medical circles about a case of a baby girl who has been “cured” or “functionally cured” of HIV with early ARV therapy.
I am not convinced. At least not as far as a real removal of all HIV from the baby is concerned.
Original abstract from CROI 2013: “Functional HIV Cure after Very Early ART of an Infected Infant”
NIH report: Toddler ‘functionally cured’ of HIV infection
Science Daily report: First Documented Case of Child Cured of HIV
Summary: a 2-year old baby, infected at or before birth, diagnosed in the first 29-30 hours on two samples, put on ARVs by 29-30 hours after birth. Viral load dropped over a month. The baby remained on ARVs for 18 months, and was off ARVs for 5 months (different reports name different drugs; the CROI abstract does not specify). Now the baby tests antibody negative and PCR negative on standard assays. On ultrasensitive PCR assays HIV can still be detected in this baby at very low levels.
The terminology: What is a cure? First of all, in the context of HIV, “cure” is a highly emotive word and should not be used unless one wishes to attract attention (or internet headlines). Second, it could be interpreted to mean the complete eradication / elimination of a disease / organism from someone’s body. That is the holy grail of HIV medicine and HIV science. When we use the word “cure” in the context of HIV, the public and media think of this. Hence the attention and headlines – many of the less scientific reports in the general press are taking it to mean this type of cure. Third, “cure” can mean attaining a state where the disease is no longer present. In the HIV context, HIV might be present but the disease is kept at bay – without ongoing ARVs. But the word “cure” is rarely used in that context – the vaccines we hope to develop for giving to HIV-infected people are not called curative vaccines. We simply hope to stop or slow down the disease process. A therapeutic vaccine that permanently halted the disease process without eradicating the HIV reservoirs from the body is not, in my experience, called a cure. (One that did, would be, but the therapeutic vaccines we are aiming for are not curative in that sense, not yet anyway.) But slowing or stopping HIV disease would be the closest thing to what has happened here – hence the use of the term “functional cure“. Since it seems clear that the virus has not been completely eradicated from this baby, I think the term “cure” is inappropriate and meant rather to get attention. What has happened is a state where the baby is infected with HIV but the HIV can do nothing to the baby. How permanent or temporary this state is for this baby is not yet known.
I will go through a few of the important issues, and discuss some of the confusion they create, and mention some of the questions they leave open.
Early diagnosis: The specificity of HIV PCR in the first few days, up to 7 days, of life is not as good as it is later. This is because infected maternal cells, and even free virus (transmitted as free virus or produced by the infected maternal cells) can enter the baby’s circulation. In most cases, such virus results in an infection in the infant, but not in all cases. Just because virus was detected in this infant so early does not mean the virus actually established a productive infection. Furthermore, in clinical grounds it’s unlikely that the infection occurred before birth – such infections usually establish themselves very well prior to birth – at best, this baby’s infection was not properly established, which means that infection must have been very recent, i.e. most likely intra/peri-partum, or during birth. Early ARVs at this stage may have been adequate post-exposure prophylaxis to prevent a productive infection. If that was the case, then one can indeed expect viral levels to drop over the next month. One would, however, expect the same if the baby were truly infected and on ARVs. What exactly happened here remains, in my opinion, debatable.
Ultrasensitive assays: The reports admit that the ultrasensitive PCRs do indeed detect low levels of HIV present in the infant. Viral DNA is currently present in cells, albeit very little (4-37 PBMCs/million PBMCs), and viral RNA is present in the baby’s blood (1 copy in the sample they took), or was at one recent point in time (free virus somewhere – which must have been produced in some cell somewhere and released into the patient’s blood.) This is probably why they are not claiming an actual sterilising cure, and are calling it a “functional cure”. However, if there is virus in the patient, then the patient is not really cured in the sense of viral eradication. The virus is there and could return with a vengeance later. Disease may indefinitely be postponed, but we currently have no idea for how long. Similarly, the Berlin patient who was cured with an allogeneic hematopoietic stem cell transplant from a donor with the CCR5-Δ32 mutation has recently been reported to have low levels of HIV in his blood [Defunct link: Science Insider, Science Mag, 11 July 2012], detectable on ultrasensitive PCR. The problem with ultrasensitive PCR (and any normal PCR) is that lab contamination can easily produce a false positive signal – as the researched cited on Science Insider says, “If you do enough cycles of PCR, you can get a signal in water for pink elephants.” So what this means is unclear and up for debate.
Absent antibodies: The reports say that the infant tests negative for anti-HIV antibodies. That is to be expected in a person who is not infected with HIV. It is also occasionally seen in infected persons, typically babies, who can no longer produce antibodies due to an overwhelming HIV infection destroying their HIV-specific immune response. This problem, and a patient we tested, are mentioned in Claassen et al., 2006; see also Garcia-Prats et al., 2012; and we identified such a baby this week in our lab. But if that were the case with this baby in the press, one would expect poor, and not good, control over her HIV, and in that case her viral load would be high, not low. I am a virologist, not an immunologist, but it seems odd to me that she a) has virus present (at very, very low levels), b) controls this virus well, and c) still has no antibodies. If she controls her virus well, it means she has a functional immune system, or at least a function anti-HIV T-cell response. In such a case I’d expect to see a decent antibody response from the B-cells induced by the T-cells. Perhaps the HIV production rate in this baby is so extremely low that the immune response can clear the minimal levels without being significantly stimulated. Certain localised viral infections that the immune system can clear or control in humans do not always result in antibody production. HPV may induce only low levels, or no levels, of antibody. Rabies patients often have no antibody detected, but that’s probably because they die before producing antibody, which they would probably do if they had lived a few weeks longer. However, while rabies is still asymptomatic (i.e. not yet causing disease up in the brain – which can be weeks to months to even years), antibodies are usually absent. The absence of anti-HIV antibodies in this baby is possibly not absolute evidence against infection, and therefore remains a debatable aspect.
So. What does this mean? What do we have? We have a baby who may or may not have been infected with HIV to begin with (assuming a real absence of virus) but tests still show detectable virus in the baby’s blood (so the baby is indeed infected unless the researchers contaminated their PCRs) and this means there was a real infection. We have a baby who has no antibodies, whatever that means in this case. We have a baby who may be controlling her virus in a spectacular way. It could mean that this baby has some genetic factor (HLA typing shows that the baby and mother shared haplotypes, but isn’t clear that the two are HLA-identical, so this may be a factor still) that allows for excellent control of her HIV (or even an ARV-assisted partial clearance and control.) Or a viral factor such as transmitted defective virus, although the maternal virus was determined to be wildtype HIV-1 subtype B – but they don’t mention any analysis on the baby’s virus, probably because there is not enough to analyse properly. We (not me, but the scientific / medical community as a whole) will need to do a lot more research on this case to work out what factors may or may not have led to this event.
My conclusion: It sounds to me as if this is a baby with a current and clearly documented HIV infection, which happens to be either extremely well suppressed and/or only minimally functional. I am of the opinion that this is not a conclusive case of a cure. The word “cure” has been qualified with the word “functional”. We should be careful using words like “cure” in situations like this – the term “functional cure” is really just the use of “functional” as a weasel word to allow the term “cure” to get into the headlines. Were this discovery presented without the word “cure”, the medics and scientists would say “Wow, excellent, we need to look at this further, this is a major breakthrough, and in time it would lead to something better than what we have now.” But since they used the word “cure”, even qualified by the word “functional”, a word the media has not understood, we have ended up with the headlines and misconceptions we see all over the internet and various other news sources.
What next? What does this mean for other babies? Many babies are put on ARVs very early. Most (i.e. all with only one known exception, and even that is inconclusive in my opinion) such babies do not clear their HIV infection. So what does this case mean for other babies? It could mean that babies do better if diagnosed very, very early. (Of course, false positives need to be screened out somehow.) It is certainly worth looking at clinical trials for treatment as early as possible, which, in spite of no clear mechanism found in this case, would be ethically justifiable based on our current knowledge. We already know that early treatment is better. Very very early treatment might well result in further cases of this sort. For this baby, I hope she is uninfected (unlikely), or will maintain life-long control. For other babies (and those with HIV who were once babies), I hope that this results in something good.
The Southern African HIV Clinician’s Society knows that all of the above are problematic, and have said the following in their statement on this case (no longer on their site). Their emphasis in bold, my comments in [red].
This is very exciting because it provides proof of concept that if an HIV infected infant is started on antiretroviral therapy early enough there is a chance that the infection can be eradicated before it takes hold in the body. [before it takes hold = before actual infection on a cellular level]
These findings do however warrant further research. [Definitely. Much more research]
This child was started on treatment exceptionally early at the age of 30 hours of life. Usually in South Africa a child is started on antiretroviral treatment following a diagnosis of HIV infection at 6 weeks. For this reason it is very unlikely that there will be any babies in South Africa in the same situation as this child. It is vitally important that children on treatment do not stop their antiretrovirals. Furthermore standard tests for HIV, conducted on children who started on treatment early on in life, may give false negative results. [They may test false negative!! This is NOT all that uncommon.] Children who have been proven to have HIV infection early on and are on treatment, should be discouraged from being re-tested later on as a result of this phenomenon.
Please don’t get irresponsible ideas from this case about how to manage your patients or your children.