Far fewer people would get PrEP in the US if generic TDF/FTC is replaced with Descovy (TAF/FTC), cost-effectiveness study finds

Doctors urged not to switch patients without clinically significant side effects
Image: Bowonpat Sakaew/Shutterstock.com

The extra price the US health system may pay for a new, branded version of PrEP, rather than an older version whose price is soon due to fall, is not remotely justifiable in terms of its marginally more benign side effect profile, a study conducted by Harvard University and presented at the Conference on Retroviruses and Opportunistic Infections (CROI 2020) has found.

Furthermore, by switching patients to the new version of PrEP “in the absence of any clinically meaningful changes in renal or bone markers”, doctors may be depriving other patients of PrEP, the authors conclude.


The study by Dr Rochelle Walensky and colleagues was published concurrently in the Annals of Internal Medicine.

Last year at the European AIDS Conference, the 96-week results of the DISCOVER study were presented. DISCOVER compared the efficacy and safety profile of two drug combinations used for PrEP: the established combination of tenofovir disoproxil fumarate (TDF) and emtricitabine (FTC) which has been marketed as Truvada, and the newer combination of tenofovir alafenamide (TAF) and emtricitabine marketed as Descovy.



In relation to medicines, a drug manufactured and sold without a brand name, in situations where the original manufacturer’s patent has expired or is not enforced. Generic drugs contain the same active ingredients as branded drugs, and have comparable strength, safety, efficacy and quality.


Cost-effectiveness analyses compare the financial cost of providing health interventions with their health benefit in order to assess whether interventions provide value for money. As well as the cost of providing medical care now, analyses may take into account savings on future health spending (because a person’s health has improved) and the economic contribution a healthy person could make to society.


Dialysis is a type of treatment that is used when the kidneys are not working properly. The dialysis machine acts as an artificial kidney, filtering excess fluid and waste products from the blood.

quality adjusted life year (QALY)

Used in studies dealing with cost-effectiveness and life expectancy, this gives a higher value to a year lived with good health than a year lived with poor health, pain or disability. 


How well something works (in a research study). See also ‘effectiveness’.

These results did show that, although the two drug combinations had equivalent efficacy and safety, there was a statistically significant decrease in kidney function in people taking Truvada compared with those taking Descovy.

The price charged by Gilead Sciences to the US Veterans’ Administration for Descovy – which probably reflects an average price – is $16,600 a year (with a maximum of about $20,000). TDF/FTC, in contrast, is either off-patent in Europe and the rest of the world, or is so de facto. In countries where it is off-patent, generic TDF/FTC has become 90-95% cheaper than the branded Truvada was, and currently costs in the region of $210-$720 a year, depending on supplier, and less when governments negotiate prices with generic companies. TDF/FTC will come off-patent in the US at the end of 2020.

Since its approval in October 2019, Descovy has captured 25% of the market for PrEP prescriptions in the US, and the manufacturer expects 40 to 45% of individuals receiving PrEP to have been switched to TAF/FTC before generic TDF/FTC becomes available.

It was therefore important to find out if the extra money that would be spent on TAF/FTC (Descovy), instead of staying on generic TDF/FTC, was worth it, in terms of the smaller number of side effects, and especially life-limiting side effects, that might be seen if PrEP users all went onto Descovy. Also to calculate, if Descovy was clearly not cost-effective in terms of the cost per fatal or disabling side effect avoided, what cost would be justified? How much would a tolerable mark-up be on the drug for its extra cost to stay within the normal limits of cost-effectiveness?

The cost-effectiveness model

Dr Walensky, who works at Massachusetts General Hospital, plus a team which included Tim Horn, Director of Medication Access and Pricing at the US National Alliance of State AIDS Directors, therefore developed a cost-effectiveness model that calculated the cost per side effect avoided if patients used Descovy for PrEP rather than generic TDF/FTC.

“We were concerned that Descovy was being pitched to doctors and patients as ‘much safer’”, Walensky told aidsmap.com, “given there has already been a massive shift to the newer drug.

“OK, we knew that the studies did show a somewhat lower occurrence of two specific side effects in patients on Descovy in some studies. There was a slight, though generally reversible, decline in kidney function in people on TDF, and there was a slight decrease in bone mineral density, and in a treatment studies, though not all, and no PrEP studies, a slight increase in fractures. What would it be worth paying to prevent these?

“For the purposes of this study we wanted to bend over backwards to be fair to TAF/FTC. So we assumed, in calculating its cost-effectiveness, that the studies had under-reported the side effects of TDF/FTC and that the reality was the absolute worst-case scenario that was remotely plausible.”

Building a worst-case scenario for Truvada

They also assumed that TAF/FTC had no side effects at all: in other words, that there simply wasn’t a downside to taking Descovy.

This is the first of a number of simplifying assumptions built into the model that we know not to be true. TAF/FTC does have its own side effects, being associated (especially when combined with integrase inhibitors such as dolutegravir for HIV treatment) with weight gain and raised cholesterol, to a greater or lesser degree.

They built a number of other assumptions into the model, some of them also counter-factual, in order to skew the study as much as possible against TDF/FTC, so they could not possibly be accused of soft-pedalling on Truvada’s side effects.

These included, for instance, assuming that every fracture caused by TDF/FTC (Truvada) was a hip fracture, as these are the most expensive to treat and have the greatest impact on health. They also assumed that, contrary to what’s been seen, the decline in kidney function seen in people on TDF/FTC was not reversible.

The model also stratified by age, and, because both bone mineral loss and kidney dysfunction are more common in older people, did a subanalysis solely of people aged over 55.

The model assumed that there were 123,610 people taking PrEP in the US, which was the actual figure estimated in late 2018.

“Our model’s findings would be no different with the larger number who are on PrEP right now,” commented Walensky. “Unless the age or gender profile had changed significantly, and it hasn’t.”

The bone-mineral density decline due to TDF/FTC was translated into fractures, as these are the main clinical consequences of loss of bone calcium. Obtained from a 2018 study of people taking TDF as treatment, these were stratified by age. This found that the risk of fracture in people under 35 taking TDF was 0.09% a year, rising to 0.3% a year in people over 54. However, not all of these were related to bone mineral loss (osteoporosis) – people of all ages have car accidents, get attacked, etc. So the risk in under-35s for osteoporosis-related fractures was 0.07% and 0.22% in over-54s.

"We assumed that the studies had under-reported the side effects of TDF/FTC and that the reality was the absolute worst-case scenario that was remotely plausible.”

This would imply a total excess of 2101 drug-related hip fractures over a five­-year time period in the 123,610 people taking PrEP if they all took TDF, compared with none if they all took TAF. They resulted in a 30% reduction in quality of life (including time off work) in the year the fracture occurred.

The consequences of a decline in kidney function were harder to calculate, partly because the worst-case scenario assumed that the decline in kidney performance on people taking TDF was irreversible and kept progressing even in people stopped TDF. This isn’t actually so.

The researchers therefore did two things. Firstly, they took actual numbers of all people with stage 2, 3 or 4 chronic kidney disease at different ages in the US, derived from the US Chronic Kidney Disease Surveillance System. They then made the counter-factual assumption that all these people would eventually progress to end-stage renal disease (ESRD) at a certain predetermined rate and would end up needing dialysis or a transplant.

They then looked at the extra cases that would be caused if all PrEP users were on TDF. Because chronic kidney disease, especially as people get older, is common, and taking PrEP is relatively rare, there would not be that many extra cases of kidney disease. It found that the incidence of chronic kidney disease in people taking TDF ranged from 0.13% in people under 25 to 3% in people over 54, but these are the proportion of people who’d eventually develop ESRD, not the proportion who’d have it immediately, as taking TDF does not cause you to develop instant ESRD. Nonetheless, these people would eventually need dialysis even if they stopped PrEP, and the associated decline in quality of life per year with ESRD would be 47% for each of the five years of the time frame of the study.

Sounds bad? Well, even with these ultra-pessimistic assumptions, the total number of excess cases of ESRD caused by 123,610 people taking TDF instead of TAF would be 25 over the course of five years.

From these assumptions, it was fairly easy to calculate the excess cost in terms of treating these two side effects of using TDF instead of TAF. The cost of treating a hip fracture was estimated as $70,400 and of a year’s worth of dialysis as $92,100 to $95,500, depending on age.

These costs were then balanced against the cost of PrEP:

  • At the full realistic cost of Descovy ($16,600 a year)
  • At a cost of half that for generic TDF/FTC once it comes off patent, so $8600.

These costs enabled Walensky’s team to then calculate the reduction in QALYs – Quality-Adjusted Life Years – if literally the only thing that reduced the quality of life of people taking PrEP was that they were taking TDF instead of TAF.

(Of note, one very important aspect of this study is that HIV incidence was not taken into account. The efficacy of PrEP was assumed to be exactly the same in people taking TDF or TAF.)

Results: QALYs, ICERs, and the budget burden

There would be an extra 690 person-years lost to death or disability in a total of 618,030 PrEP users. In other words, there would be an 0.15% maximum decrease in quality-adjusted life years if people took TDF rather than TAF – even in a scenario where TDF side effects are assumed to be worse than they actually are.

But this is not the final figure cost-effectiveness scientists like to calculate. You then have to feed in the extra cost of the new drug and ask; “Given that our new drug costs more, how much extra would we be paying for an increase in one quality-adjusted life year per individual?” This figure, the one that finally shows whether the extra cost of your new drug is worth it or not, is called the ICER – the Incremental Cost-Effectiveness Ratio.

The ICER for TAF at $16,600 a year, relative to TDF at $8300 a year, is: $7,201,200 per person.

Because over-54s have more side effects, in their case it is $3,836,700 per person. For under-35s, it’s over $16 million. That’s what it would cost extra to avoid one year’s disability in one person out of every 123,610 if you gave them Descovy as PrEP rather than TDF/FTC.

The exact threshold for cost-effectiveness is not defined in most health systems, but in the US, even for patients with non-stigmatised conditions, it is rarely above $100,000 a year (kidney dialysis being one of the rare exceptions where paying even more is thought to be worth it.)

This study shows that the cost of avoiding TDF-related side effects is far above that, even in a paper that maximises every possible TDF-related side effect and assumes TAF has none.

How much extra could Descovy cost if TDF/FTC costs $8300 in order for its ICER to come in at under $100,000? In other words, how much would a rational health system pay to avoid those hip fractures and kidney diseases cases?

"TDF going generic presents us with an amazing opportunity to get more people onto PrEP."

The answer depends on patients’ age and also on the actual discount obtained for generic TDF/FTC. But if TDF/FTC’s cost was reduced by 50% to $8300 a year, then the maximum extra price that would be worth paying for Descovy would be $8670 a year – $370 more.

In a more realistic scenario, with which the cost-effectiveness limit was $50,000, and the price of generic TDF/FTC fell by 90% as it has done in other countries, the difference would be similar: in this case TDF/FTC would cost $1660 a year, and Descovy could cost up to $1970 a year to be regarded as cost-effective – $310 more. Because over-54s have more side effects, Descovy for those patients could cost as much as $2230 a year – $570 extra.

The researchers described these mark-ups as the “maximum justifiable price” for Descovy.

The final sting in the tail of this study is to show what the cost burden of switching everyone to TAF-based PrEP would be for the US healthcare system. If the entire US HIV prevention budget of $900.8 million were spent on Descovy PrEP, it would pay for 54,300 people, or 11% of the estimated 492,000 eligible gay and bisexual men who need it (never mind other populations including women, which are thought to more than equal the gay men in number).

If TDF/FTC was available for half the price, PrEP availability could be doubled, and if it was available at 10% of the price, then it would more than meet the need for gay men. But if 45% of the US population on PrEP is switched to Descovy – then that need will never be realised.

Concluding comments

“One thing people have to realise,” comments Walensky, “is that taking a daily pill as a prevention measure is not a very common thing in medicine, and what is taken already has to be very, very safe to be ethical. TDF already probably has a more benign safety profile than, say, oral contraception.

“TDF going generic presents us with an amazing opportunity to get more people onto this effective prevention method, and especially the poorer people who need it.

“In HIV we have a history, for very good reasons, of preferring the shiny new drug over the old one, but in this case, the old pill is safe. It would be tragic if access to prevention was blocked by drug companies capitalising on the profit potential of a new pill.”


Walensky RP et al. F/TAF vs F/TDF for PrEP: how much is "better" worth? Conference on Retroviruses and Opportunistic Infections, abstract 1089LB, March 2020.

View the abstract on the conference website.

View the poster on the conference website.

Walensky RP et al. Comparative pricing of branded tenofovir alafenamide-emtricitabine relative to generic tenofovir disoproxil fumarate-emtricitabine for HIV Preexposure prophylaxis: a cost-effectiveness analysis. Annals of Internal Medicine, online ahead of print,  2020.