Simon Collins, HIV i-Base
A poster at CROI 2017 reported a case of HIV infection on PrEP with good adherence that cannot be explained by exposure to drug-resistant HIV. 
However, the case is complex with some aspects that are difficult to explain.
Previous reports of infection on PrEP have almost exclusively been related to either starting PrEP in early undiagnosed infection or to low drug levels linked to adherence. However, amongst the tens of thousands of people successfully using PrEP, two infections have so far been reported with drug-resistant HIV. [2, 3]
This new report is different for being an infection with wild-type HIV in the context of good adherence. It was actually highlighted several months ago in a press release, but with such limited details that there was little benefit from early reporting.  The poster at CROI, presented by Elske Hoornenborg from the Public Health Service in Amsterdam, now provides considerably more details.
It involves a 50-year-old man who enrolled in the Amsterdam Preexposure Prophylaxis Project (AMPrEP). He tested HIV negative at entry using both 4th generation Ag/Ab and viral load tests.
At the 8-month visit, after reporting fever for a few days, the routine Ag/Ab test was indeterminate, with atypical western blot and viral load that was undetectable. Based on a concern for resistance, PrEP was stopped and he abstained from any further sex. He was monitored with weekly viral load tests which were negative at one week, detectable at around 10,000 at two weeks – at which point he was recalled to start treatment – and viral load increased to >100,000 at the baseline sample when treatment was started.
Four-drug ART was started while waiting for drug resistance results (tenofovir/FTC, darunavir/r, dolutegravir) – and after wild-type virus was confirmed, darunavir/r was stopped. Viral load become undetectable within four weeks.
Self-reported daily adherence to PrEP was excellent, and this was supported both by pill count and high drug levels in dry blood spot (TDF-DP levels of 2234 and 2258 fmol/punch at 6 and 8 months, respectively. In the opinion of the investigator, the meticulous diary records and personal characteristics meant that less than daily adherence was not a concern.
Other factors relating to HIV risk though were very high – and these can actually be used to show the strong protection from PrEP over the previous eight months. These factors included high numbers of partners (38 to 75 per month for seven months, with a median of 3 to 5 partners on days that he had sex), frequent condomless sex, Chemsex (meth, meph, GHB, plus ketamine and coccaine – though this involved injecting on two earlier periods, this was with sterile single needles), high likelihood of mucosal trauma (duration of activity with Chemsex is generally longer with less sensitivity to trauma) and new STIs (rectal gonorrhoea twice and chlamydia).
A further detail that is difficult to explain – but actually a very important one – is the positive Ag/Ab test with not only undetectable viral load in plasma, but undetectable DNA in PBMCs and rectal samples (taken at the time of diagnosis). Generating a systemic antibody response is dependent on systemic exposure to viral load and yet these negative test show this wasn’t the case. Other examples of positive antibody responses with undetectable viral load have not been reported, especially when HIV DNA is not detected. One interpretation of this detail is that perhaps HIV infection occurred after stopping PrEP.
The authors conclude that this case highlights the importance of routine monitoring in order to detect rare cases of infection very early. The lack of resistance developing to PrEP drugs is at least one positive outcome.
As with the cases of infection on PrEP with drug-resistant HIV, this case is disappointing, but it does put PrEP now on a more realistic grounding.
Firstly, in the context of widespread PrEP use this is clearly a rare event.
Secondly, HIV infection is determined by at least a dozen various risk that each are likely to increase or decrease the risk of infection. This case is likely to be explained by an overlap of factors that are not readily measurable. These factors include volume and viral load in partners sexual fluid, type of sex, genetic factors and potentially shared injecting drug use.
It is more realistic to think of PrEP as a highly effective way to protect against HIV, supported by highest quality of evidence, even in the context of less than daily adherence.
Acceptance of some level of risk – however small – is probably a more realistic view of PrEP – as indeed it is for the high level of protection provided by Treatment a Prevention (TasP).
In the context of very large number potential exposures every year protected by PrEP, risk is clearly extremely low, whether the denominator is tens of thousands of people or millions of individual events.
- Hoornenborg E et al. Acute infection with a wild-type HIV-1 virus in PrEP user with high TDF levels. CROI 2017, 13-16 February 2017, Seattle. Poster abstract 953.
http://www.croiconference.org/sessions/acute-infection-wild-type-hiv-1-virus-prep-user-high-tdf-levels (abstract and poster)
- Knox C et al. HIV-1 infection with multiclass resistance despite preexposure prophylaxis (PrEP). 23rd CROI 2016, Boston. Late-breaker poster abstract 479aLB.
- Grossman H et al. Newly acquired HIV-1 infection with multi-drug resistant (MDR) HIV-1 in a patient on TDF/FTC-based PrEP. Research for Prevention 2016, 17-21 October 2016, Chicago. Oral abstract OA03.06LB.
- GGD Amsterdam, press release. First year Amsterdam PrEP project complete. (20 October 2016).