Comparative pricing of branded tenofovir alafenamide/emtricitabine relative to generic tenofovir disoproxil fumarate/emtricitabine for HIV pre-exposure prophylaxis: A cost-effectiveness analysis

Abstract

Background:

Tenofovir alafenamide/emtricitabine (F/TAF) was recently approved as a non-inferior potentially safer option than tenofovir disoproxil fumarate/emtricitabine (F/TDF) for HIV pre-exposure prophylaxis (PrEP) in the US.

Objective:

To estimate the greatest possible clinical benefits and economic savings attributable to F/TAF’s improved safety profile and the maximum price payers should be willing to pay for F/TAF over generic F/TDF

Design:

Cost-effectiveness analysis

Data Sources:

Published literature on F/TDF safety (with and without HIV) and the cost and quality of life impact of fractures and end-stage renal disease (ESRD).

Target Population:

Age-stratified US men who have sex with men (MSM) using PrEP

Time Horizon:

Five years

Perspective:

Health Care Sector

Interventions:

PrEP with F/TAF versus F/TDF

Outcome Measures:

Compared with generic F/TDF: fractures averted, cases of ESRD averted, quality-adjusted life years (QALYs) saved, costs, incremental cost-effectiveness ratios (ICER), and maximum justifiable price for F/TAF

Results of Base-Case Analysis:

Over a 5-year horizon, for the 123,610 MSM on PrEP, compared to F/TDF, F/TAF averted 2,101 fractures and 25 cases of ESRD with an ICER of >$7 million/QALY. At a 50% discount for generic F/TDF ($8,300/year) and a societal willingness to pay up to $100,000/QALY, the maximum fair price for F/TAF was $8,670/year.

Results of Sensitivity Analysis:

Among those >55y, the ICER for F/TAF remained >$3 million/QALY and the maximum permissible fair price for F/TAF was $8,970/year. Results were robust to alternative time horizons and PrEP-using population sizes.

Limitations:

We did not consider intermittent use or on-demand PrEP.

Conclusion:

In the presence of a generic F/TDF alternative, the improved safety of F/TAF is worth no more than an additional $370 per person per year.

INTRODUCTION

HIV prevention with pre-exposure prophylaxis (PrEP) is a cornerstone of the federal plan to End the HIV Epidemic (EHE) (1). Since the first Centers for Disease Control and Prevention (CDC) PrEP guidance in 2012, tenofovir disoproxil fumarate/emtricitabine (F/TDF) has been the only FDA-approved drug combination for PrEP use. In October 2019, based on the results of the DISCOVER Trial, tenofovir alafenamide/emtricitabine (F/TAF) became the second approved PrEP drug combination for use in men who have sex with men (MSM) as well as for transgender women (2). In the DISCOVER Trial, F/TAF was found to be non-inferior to F/TDF in terms of prevention efficacy, had comparably high drug tolerability, and showed statistically significant improvements in markers of renal and bone safety in MSM. Specifically, 48-week results for F/TAF revealed a difference in estimated glomerular filtration rate (eGFR, difference +4.1 ml/min) and bone mineral density (L-spine +1.6%, hip +1.3%) (3,4). While the manufacturer highlighted these findings as evidence that F/TAF is a superior option for people at risk for HIV who increasingly use PrEP for longer periods of time (5,6), the clinical significance of these outcomes remains a matter of debate (7).

In addition to any toxicity differences, another key consideration in decision making is the evolving cost of PrEP options; significantly less costly, generic versions of F/TDF are expected to be commercialized in late 2020 and early 2021. These potential safety, toxicity and costs trade-offs have generated an urgent call to be “forward -thinking about what should be first-line PrEP” (8).

Our analysis had the following objectives: to examine how the potentially improved safety profile of F/TAF over F/TDF might translate into observable clinical benefits; to compare the potential clinical benefits to the long-term additional costs of F/TAF; to understand the impact that the choice between these two PrEP options would have on national coverage levels, in the context of the US HIV prevention budget; and to estimate the premium that one should be willing to pay for the benefits of switching from F/TDF to F/TAF.

METHODS

Study Design and Overview

We forecast the clinical and economic outcomes of switching all MSM in the US currently receiving F/TDF-based PrEP to a comparable F/TAF-based regimen. Given the speed of new advances in HIV prevention and treatment, we adopted a short five-year time horizon. Using the methods of cost-effectiveness analysis, we sought to identify the highest possible price premium that branded F/TAF could command, even under the very best of circumstances, over generic F/TDF. Since this entailed the deliberate search for an extreme, upper-bound value, we chose extreme input values, intentionally tipping the scales in favor of F/TAF and portraying this newly approved drug in the most favorable light possible. Accordingly, we erred on the side of overstating the adverse clinical and economic consequences of F/TDF on bone and renal effects compared to F/TAF.

Population and Horizon

To determine the value of F/TAF over F/TDF for PrEP in the total population of PrEP users, we considered the age distribution of the 132,340 people currently receiving PrEP in the US (Table 1), as reported by the CDC. Because F/TAF was only FDA-approved for use in MSM and transgender women (2), we included, proportionately by age, only the 93.4% of PrEP users who are MSM (data are not available on AIDSVu for the number of PrEP users who are transgender women; Table 1) (9). Recognizing that there might be large differences in the relevant risks and benefits across different age groups, we assessed how our findings might vary when applied to specific age cohorts. We assumed PrEP would be used consistently for a five-year horizon.

Table 1.

Input parameters for analysis of the cost-effectiveness and budget impact of F/TAF vs. F/TDF

				Reference
Cohort Characteristics
2018 MSM PrEP Users (Total)	123,610			(9)
13–24 years	16,069
25–34 years	49,442
35–44 years	28,429
45–54 years	19,777
>55 years	9,888
Bone Mineral Density
Fracture risk on F/TDF, per 10,000 patient years of use
	Osteopenia	Hip	Non-hip
13–24 years	7	5	4	(10)
25–34 years	7	5	4
35–44 years	25	11	18
45–54 years	25	15	18
>55 years	22	15	15
Quality of life, in year of fracture	0.7	0.7	0.7	(12)

End Stage Renal Disease (ESRD)			Reference
Incidence, per 1,000 patient years (see methods)
	CKD 2	CKD 3–4
20–29years	73	1.3	(16)
30–39 years	113	3
40–49 years	178	11
50–59 years	207	25
60–69 years	272	55
>70 years	228	107
ESRD risk on F/TDF, per 1,000,000 patient years (see methods)
	from CKD 2	from CKD 3–4
13–24 years	43	86
25–34 years	110	220
35–44 years	254	508	(15)
45–54 years	467	934
>55 years	1196	2392
Quality of life with ESRD, for duration	0.53	0.53	(17)
Costs*
Fracture (one time)	$70,400		(18)
End-Stage Renal Disease (annual)
13–24 years	$92,100		(19)
25–34 years	$94,100
35–44 years	$93,800
45–54 years	$92,400
>55 years	$95,500
Drug Costs
Branded F/TAF	$16,600		(20)
Generic F/TDF	$8,300		Assumption, (22)

^*all costs adjusted to 2018USD

Bone Mineral Density

We obtained gender- and age-stratified data on the risk of incident fractures for patients on long-term TDF, which ranged by site from 5–25 fractures per 10,000 PY of F/TDF use (Table 1) (10). Although fracture rates were stratified by osteoporosis-related, hip, and non-hip fractures, we summed them and assumed that all fractures occurred at the hip. By choosing the fracture location associated with the largest quality of life decrement, this assumption resulted in an exaggerated 30% loss of quality-adjusted life expectancy for a full year for all patients with any type of fracture (11,12). We used other reports of F/TDF use to confirm that these fracture estimates were consistent with the goal of biasing the analysis against F/TDF-based PrEP (13).

Renal Disease

In a meta-analysis of over 9,900 patients taking F/TDF for PrEP, the majority of creatinine elevations were mild and self-limited, with only 16 and 4 subjects experiencing a Grade 2 (1.4–1.8 × ULN) or Grade 3/4 (≥ 1.9 × ULN) creatinine elevation; outcomes that were not statistically different than placebo controls (14). Further, for context, in the DISCOVER Trial, subjects on F/TAF had a baseline estimated glomerular filtration rates (eGFR) of 123 ml/min, which increased a median of 1.8 ml/min at 48 weeks; subjects on F/TDF had a baseline eGFR of 121 ml/min, which decreased a median of 2.3 ml/min at 48 weeks (4). To magnify the safety benefit of F/TAF, we assumed an accelerated and irreversible process of ESRD onset for persons receiving F/TDF. We begin the analysis with the prevalence of both CKD 2 and CKD 3–4 in the population at 0. We justify this simplifying assumption because we are only interested in cases that can be attributed to PrEP and, most specifically, in the incremental cases produced by F/TDF over F/TAF. Pre-existing cases – or cases attributable to non-PrEP causes – will cancel out in the calculation of incremental effects. From there, we used the CDC’s age-adjusted incidence for dialysis for estimated glomerular filtration rates (eGFR) 60–89 ml/min (chronic kidney disease (CKD) stage 2) and 15–59 ml/min (advanced CKD stages 3–4). For F/TDF, we then inflated the rates of progression to ESRD reported by the United States Renal Data System (USRDS) (15). Specifically, we assumed that persons with stage 2 CKD progressed to ESRD at the age-stratified rates reported for persons with CKD stage 3–4; and for those with CKD 3–4, we doubled the reported annual progression rate (16). Finally, we assumed that progression to hemodialysis was immediate and irreversible, that it lasted all five years of the treatment period and that it produced a persistent reduction in quality of life of 47% (17).

Costs

Of all fractures, those at the hip are the most expensive due to the costs of surgical repair and rehabilitation. We therefore assumed all F/TDF-related fractures had the attributable cost of a hip fracture – $70,400 (reported value, updated to 2018USD) in the year of the fracture (18). We assumed that F/TDF-related ESRD resulted in age-dependent annual hemodialysis costs, ranging from $92,100-$95,500 (reported value, updated to 2018USD) (15,19).

Currently, the Federal Supply Schedule price of F/TAF is $16,600/year (2018USD) (20); however, generic F/TDF is expected to be available in 2020 (21). Historically, prices of oral drugs fall an average of 66%, 74%, and 80% in the first 1, 2, and 5 years following the entry of a generic option (22). For generic F/TDF, we assumed a more modest price reduction of 50% from the price of branded F/TAF (to $8,300/year) that would persist over the 5-year horizon.

Additional Assumptions

We assumed equal preventive efficacy of both F/TAF and F/TDF as demonstrated in the DISCOVER Trial (3). We also assumed no excess fractures and no progression to ESRD among patients taking F/TAF. And, we deliberately excluded any quantity- or quality-of-life decrements which might arise from the reported risks of F/TAF (e.g., worsening of the lipid profile, potential increased atherosclerotic cardiovascular disease (ASCVD) risk, and weight gain) (23,24) and their associated costs (e.g., increased statin use). Given the short time horizon of the analysis (5 years), we assumed no mortality and did not apply any discounting.

Statistical Analysis

Clinical Outcomes

To estimate the expected clinical outcomes, we used fracture and ESRD risk to calculate the age-stratified per person probability of fracture or ESRD event in a given year. We then projected the expected number of fractures or ESRD events over the 5-year treatment horizon.

Cost, Cost-effectiveness, Price Premium

To calculate the total costs for the F/TDF regimen, we multiplied fracture events by per-fracture costs; age-stratified ESRD events by age-stratified dialysis costs by the 5-year horizon; and we summed these costs with that of 5-years of per person F/TDF drug costs. The cost of the F/TAF regimen included only the 5-years of per person F/TAF drug costs.

We calculated the incremental cost-effectiveness ratio of F/TAF compared to F/TDF as (Δ$/ΔQALY) between the two strategies, from the health care sector perspective. Persons receiving F/TAF were assumed to experience no quality of life decrement (quality adjustment factor = 1.0). For those receiving F/TDF, quality of life was adjusted downward to reflect the impact of fractures and ESRD (Table 1). “Willingness to pay” is a common benchmark in cost-effectiveness analysis used to measure value. While there is no generally accepted willingness to pay threshold, incremental cost-effectiveness ratios (ICER) <$100,000/QALY are widely regarded as being at the high end of acceptable in the US (25–27). ICERs below this threshold represent a reasonable return on investment; ICERs above this threshold represent insufficient return to warrant investment. Holding all other variables at their base values, we identified the price of F/TAF that would just barely achieve an ICER of $100,000/QALY, thereby identifying the highest price that payers in the US should be willing to pay for F/TAF when compared with F/TDF.

In a series of sensitivity analyses, we examined the impact of alternative costs of generic F/TDF (i.e. branded price reductions from 50–90%), of alternative willingness to pay thresholds for the F/TAF price premium, and for different treatment horizons (1–10 years).

Budget Impact Analysis

We sought to understand the potential magnitude of the financial outlays and coverage levels that could be achieved by a national PrEP scaleup campaign, using both branded F/TAF and generic F/TDF. Since our aim was to estimate the greatest possible number of eligible persons covered, we considered the extreme scenario where the entire current $900.8 million US HIV prevention budget was allocated to PrEP (28). Under this assumption, we estimated the number of people who could be provided PrEP with each regimen – inclusive of the toxicity and drug costs, and excluding HIV and sexually transmitted infection screening costs. We compared this estimated coverage to the 492,000 MSM who are projected to be PrEP eligible in the US (29).

RESULTS

Clinical Outcomes

The current population of PrEP users by age ranges from nearly 49,500 in the 24–34y age group to nearly 10,000 aged over 55y (Table 2). Accounting for both age-stratified risk and absolute numbers of patients, we estimate that an immediate switch of these persons from F/TDF to F/TAF will avert 2,101 fractures and 25 cases of ESRD over a 5-year horizon. This translates into an overall quality-adjusted life expectancy gain of 690 QALYs (F/TAF: 618,030 QALYs; F/TDF 617,340 QALYs).

Table 2.

Cumulative 5-year clinical and cost outcomes of F/TAF compared to F/TDF among MSM in the US^*

		F/TAF Outcomes		F/TDF Outcomes
Age group (years)	# on PrEP	QALYs	Costs ($)	Excess cases of ESRD	Excess fractures	QALYs	Costs ($)	ICER ($Δ/ΔQALYs)
13–24	16,069	80,344	1,333,700,000	0.2	128	80,305	675,930,000	16,960,000
25–34	49,442	247,210	4,103,700,000	2	393	247,090	2,080,600,000	16,380,000
35–44	28,429	142,150	2,359,600,000	5	764	141,910	1,235,800,000	4,681,200
45–54	19,777	98,884	1,641,500,000	7	568	98,698	863,930,000	4,163,500
55+	9,888	49,442	820,740,000	11	250	49,341	433,240,000	3,836,700
Total**	123,610	618,030	10,259,000,000	25	2,101	617,340	5,289,500,000	7,201,200

F/TAF: tenofovir alafenamide/emtricitabine; F/TDF: tenofovir disoproxil fumarate/emtricitabine; PrEP: pre-exposure prophylaxis; MSM: men who have sex with men; QALY: Quality-adjusted life year; ESRD: end-stage renal disease; ICER: incremental cost-effectiveness ratio

^*All costs reported in 2018 US dollars.

^**Values may not add to total due to rounding to five significant digits

Costs and Cost-effectiveness

Assuming F/TAF current costs ($16,600/year), F/TDF generic costs ($8,300/year), and the offsetting adverse event costs described above, switching all patients to F/TAF-based PrEP will increase total expenditures by $5.0 billion ($40,210 per person over the 5-year treatment horizon). This suggests an incremental cost-effectiveness ratio (ICER) greater than $7 million/QALY gained for F/TAF compared to F/TDF. Among patients aged over 55 years, the ICER is lower but still exceeds $3 million/QALY (Table 2). In sensitivity analyses, cost-effectiveness results of this order of magnitude persist across broad variation in the input parameter assumptions, including: treatment horizons ranging from 1 to 10 years (base case 5 years); PrEP-using population sizes ranging from 100,000 to 1 million (base case 123,610); and focusing only on the oldest, highest-risk patients. Relaxing any of the pessimistic assumptions regarding F/TDF’s renal and bone safety, (for example, assuming similar safety profiles for both treatments), only makes F/TAF even less cost-effective.

Maximum Justifiable Price for F/TAF

Using, as a point of departure, a societal willingness-to-pay threshold of $100,000/QALY (25–27) and a generic F/TDF cost of $8,300/year (50% generic cost reduction), the superior safety profile of F/TAF, based on the assumptions above, could justify a price of up to $8,670/person/year for the overall PrEP population; at a willingness-to-pay threshold of $150,000/QALY and an F/TDF cost of $1,660/person/year (90% generic cost reduction) the highest justifiable cost of F/TAF would be $2,090/person/year (Table 3). For patients aged >55 years, slightly higher F/TAF prices could be justified ($8,970/person/year and $2,430/person/year, respectively) due to the higher risks of adverse events with F/TDF (Table 3).

Table 3:

Permissible cost of F/TAF under alternative willingness-to-pay thresholds and F/TDF costs (generic discounts)

	Willingness to Pay Threshold
Annual F/TDF cost (% price reduction)	<$50,000/QALY	<$100,000/QALY	<$150,000/QALY
Total Population
$8,300 (50%)	$8,610	$8,670	$8,730
$4,150 (75%)	$4,460	$4,520	$4,580
$1,660 (90%)	$1,970	$2,030	$2,090
Population >55 years
$8,300 (50%)	$8,870	$8,970	$9,070
$4,150 (75%)	$4,720	$4,820	$4,920
$1,660 (90%)	$2,230	$2,330	$2,430

Budget Impact Analysis

If the entire US budget for HIV prevention ($900.8 million) were devoted to PrEP, a nationwide rollout using branded F/TAF ($16,600 per person per year) could achieve a coverage level no greater than 54,300 (or 11%) of the estimated 492,000 eligible MSM. This coverage level could be doubled (quadrupled) by switching to a generic F/TDF alternative priced at a 50% (75%) discount to the branded option.

DISCUSSION

It is estimated that 1.2 million Americans are at risk for HIV and eligible for PrEP (29). At the current Federal Supply Schedule Price of $16,600/year (2018USD) for branded F/TAF (20), a nationwide PrEP program using this agent would consume the entire $900.8 million federal budget for HIV prevention several times over (28). The relevance of a soon anticipated generic F/TDF option will be apparent to insurers and other payors who must decide whether the improved safety of F/TAF justifies its premium price (30). Patients and providers may need to be convinced that this also matters to them: that higher costs may decrease both access and long-term adherence to PrEP; and that the risks of attrition due to higher costs need to be weighed against any safety benefits of switching to F/TAF.

The DISCOVER Trial demonstrated statistically significant differences between F/TAF and F/TDF in surrogate markers of bone and renal safety (3). We sought to quantify how those differences in intermediate markers might translate into meaningful clinical outcomes, in order to understand the magnitude and relevance of these outcomes and how much payers, and society more broadly, should be willing to pay for them. Using the methods of cost-effectiveness analysis with input parameter assumptions that cast F/TAF in a highly favorable light, we found that a markup for F/TAF of up to $370 over the price of generic F/TDF ($8,300) could be justified based on those toxicity differences. This implies that, in the presence of a generic alternative, the current price of F/TAF ($16,600/year) would have to be reduced by over $7,900 per year for F/TAF to satisfy generally accepted standards of societal value. If F/TDF can achieve the 75% price reduction that is commonly observed when generic competition ensues (i.e., a cost of $4,150 per year), the F/TAF price would need to be no higher than $4,520 to demonstrate value based on cost-effectiveness. For older patients at unusually high risk of renal disease or bone-related adverse events, the switch from F/TDF to F/TAF would have greater clinical impact and benefit. Even in this population, however, it would be difficult to defend a price of greater than $800 over the cost of the generic alternative.

We also find that it is the cost of drug – more than the frequency or cost of adverse events – that is likely to limit PrEP use and access. At half the drug cost of F/TAF, PrEP with F/TDF could cover about twice the number of people on any given budget. Because DISCOVER demonstrated that incidence is four-fold higher among people not taking PrEP compared to people taking it, excess drug costs, resulting in decreased PrEP coverage, potentially translates into the lost opportunity to prevent tens of thousands of new HIV infections.

For almost a decade, tenofovir disoproxil fumarate/emtricitabine (F/TDF) has been a mainstay of HIV prevention and has been repeatedly shown to be safe, effective, and cost-effective when used as PrEP in high-risk populations (14,31,32). The economic value of F/TDF will increase, as emtricitabine loses patent protection and co-formulated generic alternatives becomes available. F/TDF’s successful track record of prevention explains why the recent approval of F/TAF for PrEP has generated such important discussion. Since its approval in October 2019, F/TAF has captured 25% of the market for PrEP prescriptions, and the manufacturer expects 40–45% of individuals on PrEP to have been switched to F/TAF before generic F/TDF becomes available (33). Anecdotal evidence suggests that many providers are actively switching their PrEP patients to the newest available option, F/TAF, in the absence of any clinically meaningful changes in renal and bone markers while on F/TDF. A recent study from the Swiss cohort finds that switches to F/TAF for HIV treatment occurred in over 50% of patients without any indication for change (34). Investigators in that study cited risks of TDF toxicity as a reason to change the regimen, including: eGFR <60 ml/min, marked proteinuria, or osteoporosis (as measured by T-score or fragility fractures). Payers are also contemplating who might benefit by initiating PrEP with, or switching to, an agent that has not been shown to be more effective in preventing HIV infection and that offers statistically significant but very small improvements in surrogate markers of safety, at brand-name prices.

Interpretation of this work should be considered in the context of its limitations. First, we remind readers that each assumption and uncertain data parameter underlying our analysis was selected to cast F/TAF in the best possible light. In addition to presenting extreme values for bone and renal effects beyond those that have ever been reported – both from a clinical and cost perspective – we intentionally excluded the potential beneficial effects of F/TDF compared to F/TAF. While hypertension and diabetes are often cited as reasons to favor the renal benefits of F/TAF, mounting evidence suggests the comparatively detrimental effects of F/TAF on lipid profiles, weight gain and ASCVD risk (23,24). Had we chosen to include these effects in this analysis, the justifiable premium for F/TAF would have been reduced even further. Second, while we considered a five-year PrEP horizon and daily, rather than on-demand PrEP, we recognize that PrEP consumers may come on and off PrEP based on need and may choose on-demand options. To the extent that adverse events and drug costs are similarly proportional to use, we believe our findings are generalizable to those circumstances. Third, pricing variables are based on FSS prices available to federal purchasers, which may differ from list prices and prices net of the highly variable discounts or rebates that accrue to Medicaid, commercial insurers, and safety net providers. Fourth, this analysis does not address barriers to PrEP beyond drug costs, such as stigma, insurance coverage, or other social determinants, nor does it account for public and private programs subsidizing PrEP drug access for uninsured individuals. Finally, we highlight that these findings may not be generalizable to other PrEP-using populations, including cis- or transgender women and people who inject drugs.

PrEP is a key pillar of the US plans to End the HIV Epidemic. Generic F/TDF could greatly expand PrEP coverage among the most price-sensitive members of the at-risk population. If branded F/TAF drives out generic F/TDF and inhibits acceptability, access and uptake, overall rates of PrEP coverage in the at-risk population could fall, and F/TAF could end up causing more avoidable HIV transmissions than it prevents.

F/TDF for PrEP is effective, cost-effective and safe; soon, it will also be much less expensive. At current F/TAF prices, F/TAF compared to generic F/TDF will not be cost-effective in the US, even in populations at highest risk for F/TDF adverse events. Given the very small, albeit statistically significant, differences in surrogate markers, without evidence of clinical significance, there is no urgency and no reason to switch PrEP regimens now, and it would be hard to switch back later. In about a year, when F/TDF is generically available, payers should consider the $370 premium ceiling estimated here in assessing whether to recommend that patients switch to F/TAF.