Two studies that compared intracellular levels of the drugs used in PrEP found that levels of tenofovir in cells were seven times higher when the newer version of the tenofovir prodrug, tenofovir alafenamide (TAF), was used, than when the older version, tenofovir disoproxil fumarate (TDF), was used.
TDF has been used for HIV treatment for 20 years and is a component of most PrEP tablets, including Truvada. Tablets containing TAF have only been marketed in the last five years and include Descovy, which has been licensed for use as PrEP in the United States, Canada, Australia and Taiwan, but not elsewhere.
Both TDF and TAF are ‘prodrugs’. A prodrug is the formulation of a drug that needs to be given to the patient in order for it to get into the body, e.g. because it can be absorbed by the gut, but which then needs to be metabolised by the body to turn it into the active drug – in this case, tenofovir disphosphate (TFV-DP). TAF was developed in response to concerns that the slow metabolism of TDF meant that high levels of it persisted extracellularly in the blood, and that the well-known side effects of kidney dysfunction and bone demineralisation were attributed to this.
The intracellular concentration of tenofovir disphosphate reached when TAF was taken once every three days was 2.6 times higher inside peripheral blood mononuclear cells (PBMCs) than when TDF was taken daily. PBMCs are the lymphocyte and monocyte cells in the immune system that are floating freely in blood and thus easy to extract; PBMCs include some of the T-cells which HIV mainly infects.
The version of tenofovir used did not have a significant impact on the other drug currently included in PrEP pills, emtricitabine (also called FTC).
Two separate small studies, both conducted by Professor Peter Anderson’s group at the University of Colorado, looked at drug concentrations in people at low risk of HIV who took PrEP pills either daily, twice in three days, or once in three days (i.e. at 100%, 67% and 33% of daily dosing).
They took Directly Observed Therapy (DOT) either in person or via video link. The two studies, one for TDF and one for TAF, involved volunteers taking one of the three dosing regimens for 12 weeks, having a 12-week ‘washout’ period off drug, and then taking another one of the three regimens for 12 weeks – so each volunteer took two of the three dosing regimens.
There were 44 volunteers in the study of TDF/FTC: 32 volunteers took a daily dose, 16 took the 67% dose, and 15 took the 33% dose. In the TAF/FTC study, there were 35 volunteers and the three dosing regimens were each taken by 23 volunteers.
The makeup of the study population in the two studies was largely similar, with one important exception. In both studies the average age of participants was 29 and the gender balance was roughly 50/50. They were of average weight (74kg) and their kidney function was excellent at 120 millilitres per minute (ml/min) creatinine clearance.
However there were fewer non-White people in the TAF study than in the TDF study. This may reflect the fact that while the TAF study was solely conducted in Colorado, the TDF study was a collaboration with the San Francisco Department of Public Health and was partly conducted in that more racially mixed city. In the TDF study, 57% of participants were White and of those, nearly half were Hispanic, whereas in the TAF study, 83% were White and of those, 20% were Hispanic. There were more Black and Asian participants in the TDF study.
This matters because in the TDF study, intracellular TFV-DP levels turned out to be significantly lower in participants who were not White.
As expected, TFV-DP levels were considerably higher in the TAF recipients. Measured in femtomols per million PBMCs (fmols/106 cells), the levels achieved with 100% dosing were 593 for TAF versus 82 for TDF; for 67% dosing 407 versus 57; and for 33% dosing 215 versus 32.
Levels of the companion drug FTC (or rather its intracellular metabolite, FTC-TP) were not affected by the tenofovir prodrug used. For TAF and TDF respectively, they were 8000 versus 7200 fmols/106 cells for 100% dosing, 6100 versus 5500 for 67% dosing and 4000 versus 3600 for 33% dosing.
"The real gap in knowledge is that researchers have never evaluated TAF’s efficacy with event-based dosing."
What this means is that for 100%, 67% and 33% dosing, levels of intracellular TFV-DP were respectively 6.7, 7.1 and 7.3 times higher in TAF than in TDF recipients. Perhaps more significantly, levels of TFV-DP were 2.6 times higher in TAF recipients taking only one pill every three days than in TDF recipients taking one pill every day.
The half-lives of TFV and FTC in cells – the rate at which the drugs left the body – was roughly similar, though a bit longer for TAF at 69.6 hours, compared with 54.5 hours for FTC, and 50.4 hours for TDF.
Lower kidney function led to somewhat higher drug levels; this is as you’d expect because the kidneys were processing the drugs less efficiently. For every 10 ml/min decline in creatinine clearance, TFV-DP levels rose by 2% and FTC-TP levels by 5%. Given the generally healthy kidney function of study participants, however, we can’t predict what drug levels might be in people with poorer kidney function.
As mentioned above, on TDF but not on TAF, non-White participants had intracellular TFV-DP levels that were 28% lower than in White participants, and FTC-TP levels 13% lower. However because the two studies were not similar in terms of ethnic makeup, it’s hard to generalise this finding.
Age and gender did not affect intracellular drug levels in either study.
Implications for PrEP and for future research
What this study does not show is that TAF has better efficacy than TDF when used as PrEP. This is an implication, but we have knowledge gaps about precisely how PrEP works. Because less TDF gets into cells, extracellular levels of it in blood plasma and on mucous membranes can be higher than levels of TAF – and this may be as important to efficacy as intracellular levels.
When used as treatment for people with HIV, studies have found slightly higher viral suppression rates in patients taking TAF, but only in patients also taking the boosting agents ritonavir or cobicistat, i.e. in patients also taking protease inhibitors or elvitegravir.
When used as PrEP, we’ve known since the iPrEx study ten years ago that intracellular levels of tenofovir do predict PrEP efficacy. But that study only used TDF.
Two years ago the randomised study comparing TDF or TAF as part of PrEP, DISCOVER, found that HIV incidence in people taking TAF was half what it was in people taking TDF – but because both drugs were so good at preventing HIV, this difference was not statistically significant. A post-hoc analysis did find a suggestion that HIV infection was more likely in people taking 2-4 TDF pills a week than in people the same number of TAF pills. This suggests that TAF could, in theory, be more ‘forgiving’ of relatively low rates of adherence to PrEP than TDF.
The real gap in knowledge, however, is that researchers have never evaluated TAF’s efficacy when lower-than-daily rates of PrEP usage are not due to ‘poor adherence’ - but are part of a predetermined dosing strategy such as event-based or 2-1-1 dosing.
Event-driven PrEP has only ever been evaluated in one quite small randomised controlled study that used TDF/FTC. It has never been licensed as an indication by the US Food and Drug Agency, though pragmatic use of this strategy is now on the increase in the US, and population studies indicate its efficacy is good.
Given these increasing rates of use, and the fact that 2-1-1 PrEP has been endorsed by the World Health Organization, it would seem important to directly compare the efficacy of TAF and TDF when both are used in event-based dosing. Without doing that, we can’t find out whether, in event-driven PrEP, the greater cost of TAF might be compensated for by greater efficacy.
Yager JL et al. Tenofovir-diphosphate in peripheral blood mononuclear cells during low, medium, and high adherence to F/TAF vs. F/TDF. AIDS, published online ahead of print, 2 September 2021.