modelling the budget impact of darunavir in the treatment of highly treatment-experienced,...

Modelling the Budget Impact of Darunavirin the Treatment of Highly Treatment-Experienced, HIV-Infected Adultsin FranceXavier Colin,1 Antoine Lafuma,1 Dominique Costagliola,2 Erik Smets,3 Josephine Mauskopf 4

and Pascal Guillon5

1 Cemka-Eval, 43 Boulevard du Marechal Joffre, 92340 Bourg-la-Reine, France

2 INSERM, Unite Mixte de Recherche (UMR) S 720, and Universite Pierre et Marie Curie-Paris 6,

UMR S 720, Paris, France

3 Johnson & Johnson Pharmaceutical Services LLC, Mechelen, Belgium

4 RTI HEALTH Solutions, Research Triangle Park, North Carolina, USA

5 Janssen Cilag SAS, Issy les Moulineaux, France

Abstract Background: A key element for payers in the assessment of the economic

profile of a medication is its anticipated impact on the evolution of healthcare

budgets.

Objectives: To forecast the impact of the use of darunavir with low-dose

ritonavir 600/100mg twice a day (bid) in highly treatment-experienced, HIV-

infected adults who have failed one or more protease inhibitor (PI)-containing

regimen on the budget of the French Sickness Fund (French healthcare sys-

tem) over a 3-year time horizon.

Methods: A transition state model based on disease severity was developed

that compared the evolution of antiretroviral and non-antiretroviral-related

direct costs of care in the target population over 3 years (2007–2009) under

two scenarios: (1) darunavir enters the French market in year 1; (2) darunavir

is not available to the target population during 2007–2009.Model inputs were

derived from a targeted analysis of the French hospital database in HIV, the

darunavir POWER 1 and 2 trials and other relevant clinical studies.

Results: In the scenario where darunavir was available from year 1, the pro-

portion of patients in the lower, more costly CD4 cell count strata (£100 cellsper mm3) was consistently lower than in the scenario without darunavir in

each year of the model (17.0% vs 19.2%, 13.9% vs 18.3% and 10.8% vs 16.8%for years 1, 2 and 3, respectively). As a result, over the entire 3-year period,

the net increase of antiretroviral drug costs (+5.6 million Euros; h), resulting

from the substitution of older, cheaper PIs by darunavir, is expected to be

fully compensated by savings in hospitalization costs (h-9.7 million) and

ORIGINAL RESEARCH ARTICLEPharmacoeconomics 2010; 28 Supp.1: 183-197

1170-7690/10/0001-0183/$49.95/0

ª 2010 Adis Data Information BV. All rights reserved.

expenditures for otherHIV-related (non-antiretroviral) medications (h-7.3million),

leading to a net saving of h11.4 million or 2.9% of the total budget in the

scenario without darunavir. Various sensitivity analyses confirmed these

projected savings.

Conclusion: The use of darunavir/ritonavir (DRV/r) 600/100mg bid, in com-

bination with other antiretroviral agents, in highly pre-treated, HIV-infected

adults who have failed one or more PI-containing highly active antiretroviral

therapy regimen is not expected to increase the budget of the French

healthcare system, in comparison with a scenario without darunavir. Further

research is needed to estimate the budget impact of the use of DRV/r in less

treatment-experienced, HIV-infected individuals in France.

Introduction

The greater life expectancy of HIV/AIDSpatients associated with the use of highly activeantiretroviral therapy (HAART) has resulted inan increase in the total lifetime costs of HIV in thedeveloped world, where HAART is now widelyavailable.[1-4] HAART has also shifted the driversof HIV-related expenditures from hospital in-patient costs to the pharmacy and outpatientsettings, causing a sharp reduction in hospitali-zation-related costs that has partly offset the in-creased HAART costs.[5-8]

One of the main challenges in HIV care is thesuccessful management of patients who have failed,and often become resistant to, currently availableantiretroviral agents, in particular protease inhibi-tors (PIs), which are the cornerstone of HAART intreatment-experienced individuals.[9-12] Several newantiretroviral drugs for people who have failed ex-isting HAART regimens have recently been ap-proved, including the PI darunavir.[13-17]

As part of its ongoing clinical developmentprogramme, darunavir has been evaluated intreatment-experienced, HIV-infected adults whohad failed more than one PI-containing HAARTregimen in two independent, randomized, con-trolled, multinational, phase IIb trials (POWER 1and 2; TMC114-C213 and C202; PerformanceOf TMC114/r When evaluated in treatment-Experienced patients with protease inhibitor Re-sistance).[18] The POWER trials, which supportedthe initial accelerated/conditional approval ofthis drug, demonstrated that darunavir combined

with low-dose ritonavir (DRV/r, 600/100mgtwice a day; bid) and an optimized backgroundregimen (OBR) was well tolerated and generateda significantly superior virological response after48 weeks of therapy compared with currentlyavailable control PIs.[13,14,18] In addition, meanchanges from baseline in CD4 cell count weresignificantly greater for DRV/r than for controlPIs at all timepoints.

Elsewhere in this supplement, several authorshave demonstrated the cost effectiveness of DRV/r600/100mg bid in highly treatment-experiencedpatients.[19-21] Another important element in the as-sessment of the economic profile of a medicationfrom the payer perspective, however, is its antici-pated impact on the evolution of healthcare budgets.

We, therefore, developed a budget impactmodel to analyze the anticipated effect on theFrench Sickness Fund over a 3-year time horizonof the introduction and use of DRV/r 600/100mgbid in highly treatment-experienced, HIV-infectedadults who have failedmore than one PI-containingregimen and are eligible for treatment withdarunavir-based HAART. The French SicknessFund perspective is similar to that of the Frenchhealthcare system, as HIV-infected patients arefully reimbursed for their healthcare expenses.

Methods

Model Structure and Main Assumptions

The overall design of the model used for thecalculation of the budgetary impact is shown in

184 Colin et al.

ª 2010 Adis Data Information BV. All rights reserved. Pharmacoeconomics 2010; 28 Supp.1

figure 1. This transition state model takes intoaccount disease severity expressed as CD4 cellcount levels. The main assumptions driving themodel were that: (1) treatment of the targetpopulation with darunavir-based HAART willresult in a substantially greater improvement ofHIV-induced immune deficiency (measured byCD4 cell count), compared with other currentlyavailable PIs and that (2) different degrees ofHIV-induced immune deficiency are associatedwith different levels of healthcare expendituresin general, and non-antiretroviral-related costs inparticular. Six discrete CD4 cell count ranges,chosen on the basis of their clinical relevance,were postulated in the model and attributedan annual patient cost: A (£50 cells/mm3);B (51–100 cells/mm3); C (101–200 cells/mm3);D (201–350 cells/mm3); E (351–500 cells/mm3)and F (>500 cells/mm3).[22]

A target population was identified in the modeland assigned a given baseline steady-state CD4cell count distribution. Clinical trial data fromthe DRV/r 600/100mg bid arm of the pooledPOWER 1 and 2 trials over 48 weeks of therapywere then used to estimate the shifts in thisbaseline CD4 cell count distribution associated

with the use of DRV/r.[18] Based on the entryCD4 cell count distribution and the size of theshift upwards, the model calculated the projecteddistribution of the target population among dif-ferent CD4 cell count ranges over time in twodifferent scenarios/treatment strategies:1. One scenario/strategy, in which all patientseligible for DRV/r continue to be treated duringthe time horizon of the model with an evolvingmix of HAART regimens containing currentlyavailable PIs other than darunavir, includingtipranavir (another ritonavir-boosted PI that hasbeen evaluated in highly treatment-experiencedpatients),[23] together with standard OBR with orwithout the fusion inhibitor, enfuvirtide.2. An alternative strategy in which darunavirbecomes available for eligible patients in year 1and an increasing proportion of this targetpopulation across the model time horizon receiveHAART combinations, consisting of anOBRwithor without enfuvirtide and DRV/r 600/100mgbid, in lieu of other currently available PIs, inline with the currently approved indication fordarunavir.

These distributions were calculated as weight-ed means based on the proportion of patients

ART mix inscenario

without DRV

ART mix in scenariowhere DRV

increasingly used

Study sample(DRV target population)

Resources consumed perCD4 cell count range

Estimated budget impact ofDRV over the model

time horizon

Costing• ART costs• Other treatment costs• Hospitalization costs

CD4 cell countdistribution

CD4 cell countdistribution

Overall costs ofscenario without DRVduring year 1, 2 and 3

Overall costs ofscenario with DRV

during year 1, 2 and 3

Fig. 1. Schematic representation of the budget impact model of darunavir. ART = antiretroviral therapy; DRV =darunavir.

Budget Impact of Darunavir in Highly Experienced Patients 185


receiving DRV/r or other currently availablePIs. Finally, this projected CD4 cell count dis-tribution allowed calculation and comparison ofthe antiretroviral-related and non-antiretroviral-related direct costs associated with the evolutionof the antiretroviral therapy (ART) mix over timein the two different scenarios during each year ofthe model time horizon, thus estimating thebudget impact associated with the anticipated useof DRV(/r).

Model Perspective

The economic perspective was that of the FrenchSocial Security system, in which the reimburse-ment rate for healthcare expenses of patients withHIV is close to 100%. The study period for themodelling covered 2007–2009, which correspondsto the 3-year period for which sales estimates aretaken into account by the French health author-ities for price negotiation at the time of marketingof a new drug.[24]

Modelled Target Population

The patient sample considered by this modelcorresponded to the target population of patientsfulfilling the indication for darunavir approvedby European regulators in February 2007.[14] Thesize of the sample was estimated by an analysis ofthe French Hospital Database in HIV (FHDH).[25]

The FHDH is one of the largest cohorts of HIV-

infected patients in the world, accounting for50–60% of HIV-treated patients in France, andcollects prospective data on HIV-infected in-dividuals seeking care in 62 specialized Frenchteaching hospitals. By mid-2005, 109 499 patientswere represented in the database. The patientsample included in the analysis of FHDH datacorresponded to HIV-infected individuals with atleast two visits during the study period 2003–2005and with at least one CD4 cell count available;they were receiving HAART and had at leastone documented PI failure. Treatment failurewas defined as an HIV-1-RNA level greater than1000 copies/mL and two previous visits withHIV-1-RNA levels of less than 500 copies/mL(secondary failure), or HIV-1-RNA levels greaterthan 500 copies/mL 6 months after treatment in-itiation (primary failure).

In the FHDH database, 44 819 HIV patientswere identified with at least two visits and oneCD4 cell count during the study period (figure 2).Among this population, 37 521 patients receivedHAART and 2080 of these had at least onedocumented PI failure and were currently failingtheir most recent PI treatment, therefore meetingthe criteria of the model. These 2080 patientsrepresented 4.6% of the HIV population includedin the FHDH database.

By extrapolating this figure to the total num-ber of patients with HIV infection in France, es-timated at 130 000 patients in 2006 by the French

FHDH database National population

All identified patientsn = 44 819

Treated with HAARTn = 37 521

Patients with ≥1prior PI failure n = 2080

All patients with HIVn = 130 000

Target populationn = 5980

4.6%

Experiencing (new) failure

Fig. 2. Construction of the study sample. FHDH =French Hospital Database in HIV; HAART =highly active antiretroviral therapy;PI = protease inhibitor.

186 Colin et al.


Ministry of Health,[26] the total number ofpatients eligible for treatment with darunavir inFrance was estimated at 5980 patients in 2006.

Incremental increases in the study sample sizeover time were estimated using data on the emer-gence of PI failure from the FHDH database.Among the 2080 patients fulfilling the treatmentcriteria, 756 presented with a first PI failure afteran average exposure period of 3.2 years, with anannual death rate of 4.7%. The incremental increasewas thus calculated as (756/2080)/3.2–0.047, cor-responding to an annual increase of 6.6%, yield-ing a predicted target population of 6375 HIV/AIDS patients in 2007, year 1 of the model. Fur-ther growth of the study sample size over time ispresented in table Ia.

Evolution of the Treatment Mix Withand Without Darunavir

Data on the proportional use of darunavir andthe other currently available PIs in the targetpopulation in the two modelled scenarios (withand without darunavir) were based on marketforecasts provided by the manufacturer of dar-unavir, an analysis of FHDH data, a review ofthe recent sales of antiretroviral drugs in France

and clinical expert opinion. Table Ib shows theproportion of the target population and thecorresponding patient numbers that were anti-cipated to be treated with DRV/r and other cur-rently available PIs during the different yearsof the model, in a scenario with and withoutdarunavir.

In the POWER 1 and 2 trials, tipranavir/ritonavir (TPV/r), was not included in the controlarm because it was not commercially available atthe time the studies were recruiting.[18] TPV/r has,however, demonstrated efficacy in the RESIST 1and 2 (Randomized Evaluation of StrategicIntervention in multi-drug reSistant patients withTipranavir) trials in a highly pre-treated, PI-resistant patient population that was broadlycomparable to that of POWER 1 and 2.[23,27]

Therefore, in the scenario without darunavir, itwas estimated that TPV/r would be given to anincreasing proportion of HIV-infected in-dividuals, otherwise eligible for DRV/r, as a re-sult of the progressive failure and discontinuationof currently available PIs and the need for newtreatments.

Because of the more favourable risk/benefitratio and pharmacokinetic interaction profile ofdarunavir, its uptake during the time horizon of

Table Ia. Study sample size

2006 2007 2008 2009

Study sample size (darunavir target population) 5980 6375 6796 7245

Incremental increases in the study sample size over time were estimated using data on emergence of protease inhibitor failure from the French

Hospital Database in HIV.[25]

Table Ib. Anticipated market shares of darunavir, tipranavir and other currently available protease inhibitors over time in two scenarios:

darunavir enters the French market in year 1 and darunavir does not enter the market

Darunavir enters the market in year 1 Darunavir does not enter the market

2006 2007 2008 2009 2006 2007 2008 2009

Market share of darunavir (%) 0 20 45 70 0 0 0 0

Darunavir-treated patients (estimated) 0 1275 3058 5071 0 0 0 0

Market share of tipranavir (%) 5 5 5 5 5 10 22 46

Tipranavir-treated patients (estimated) 299 319 340 362 299 638 1495 3333

Market share of other PIs (%) 95 75 50 25 95 90 78 54

Other PI-treated patients (estimated) 5681 4781 3398 1811 5681 5738 5301 3912

Data on the proportional use of darunavir and the other currently available PIs in the target population in the two modelled scenarios (with and

without darunavir) were based on market forecasts provided by the manufacturer of darunavir, an analysis of FHDH data, a review of recent

sales of antiretroviral drugs in France and clinical expert opinion.[24-26]

FHDH =French Hospital Database in HIV; PIs =protease inhibitors.



the model in a scenario in which darunavir wouldbecome available for the treatment of the targetpopulation in France in year 1 was estimated tobe higher than the forecasted uptake of tipranavirin a scenario without darunavir in the base-caseanalysis.[18,23,28] Nevertheless, in the base case,TPV/r was still expected to be used in a small, butstable proportion of the target population in theevent that darunavir was to become available.

CD4 Cell Count Distribution of the TargetPopulation With and Without Darunavir:Baseline and Evolution with Time

The baseline distribution (table II) of the tar-get population across the CD4 cell count strataconsidered by the model at the start of the modeltime horizon was derived from the analysis ofFHDH data for patients having at least one PIfailure. As eligible participants were receivingtreatment with available antiretroviral drugs atthe time of study initiation, the model assumedthat the baseline CD4 cell count distribution wasin steady state in patients eligible for treatmentwith DRV/r who were receiving currently avail-able PIs other than TPV/r (plus OBR).

In the pooled POWER 1 and 2 trials, DRV/rwas associated with a mean increase in CD4 cellcount from baseline of 92.4 and 102.0 cells/mm3

at 24 weeks and 48 weeks, respectively (p < 0.0001vs control; standard deviations 113.6 and126.94).[18,27] In the pooled RESIST trials, TPV/r

was associated with a mean increase from base-line of 45 CD4 cells/mm3 (p < 0.001 vs control) at48 weeks.[23,27] The model, therefore, assumedthat these PIs would lead to an upward shift inthe baseline steady-state CD4 cell count dis-tribution, based on the relative use of TPV/r andDRV/r in the two scenarios (with and withoutdarunavir). In the calculations of this shift atdifferent timepoints (years 1, 2 and 3), the modelassumed that all patients in each CD4 cell countstratum had a CD4 cell count equal to the mid-point of the range. The new distribution ofpatients within the CD4 cell count strata wascalculated by shifting these midpoints up ac-cording to a normal distribution using the meanand standard deviation of the increase in CD4 cellcount generated by DRV/r (and TPV/r). Year 1and 2 CD4 cell count distribution shifts werebased on the incremental CD4 cell count changesfrom baseline observed after 24 and 48 weeks oftherapy, respectively. The model assumed that inyear 3, no further increases in CD4 cell countwere experienced by the target population.

For example, in the stratum of patients with abaseline CD4 cell count less than 50 cells/mm3,patients were assumed to have an average CD4cell count of the midpoint of 25 cells/mm3 atbaseline. At year 2, the modelled incremental in-crease in CD4 cell count in these patients treatedwith DRV/r was 102 cells/mm3, yielding an aver-age CD4 cell count of 127 cells/mm3. A normaldistribution was then applied around this new

Table II. Distribution of patients between CD4 cell count ranges at baseline and during years 1, 2 and 3 according to treatment (optimized

background regimen, darunavir/ritonavir+OBR and tipranavir/ritonavir+ optimized background regimen)

CD4 cell count OBR DRV/r +OBR TPV/r +OBR

Baseline and years 1, 2 and 3 (%) Year 1 (%) Year 2 (%) Year 3 (%) Year 1 (%) Year 2 (%) Year 3 (%)

0–50 11.2 3.2 3.3 3.3 7.1 6.8 6.8

51–100 8.7 4.1 3.8 3.8 5.8 6.3 6.3

101–200 19.5 14.2 13.3 13.3 16.1 17.0 17.0

201–350 28.1 27.8 27.2 27.2 27.7 28.3 28.3

351–500 19.0 24.6 24.8 24.8 22.0 21.6 21.6

>500 13.5 26.2 27.6 27.6 21.4 20.1 20.1

Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0

Data on baseline distribution were obtained from the FHDH database;[25] the final distribution was estimated by applying the proportions

shifting between CD4 cell count ranges because of the improved efficacy. No change is assumed in OBR from baseline to year 3.

DRV/r =darunavir/ritonavir; FHDH =French Hospital Database in HIV; OBR = optimized background regimen; TPV/r = tipranavir/ritonavir.

188 Colin et al.


average CD4 cell count to estimate the new dis-tribution of patients in this group treated withDRV/r across the different CD4 cell count stratain the model at year 2.

Finally, a weighted average was applied ineach year of the model to calculate, for each ofthe two scenarios, the new distribution of thetarget population across the different CD4 cellcount strata, as a function of the relative use ofDRV/r, TPV/r and other PIs in years 1, 2 and 3,and the impact of these individual drugs on theCD4 cell count distribution (detailed in table II).

Costs Considered by the Model

All costs were reported in Euros and no ad-justment was made for inflation (because a largeproportion of the HIV-related costs are due tomedications, and fixed prices are negotiated atmarket entry). Costs were based on the latestavailable information. Medication, consultationand visit costs were based on 2008 prices. For theinpatient costs, we used the last available nationalcosts scales (2005) to provide true costs perdiagnosis from a national sample of French hos-pitals. This scale was then used for establishingthe tariffs for 2008.

Darunavir and tipranavir costs were the 2008prices, with a daily cost of h24.84 for darunavir

and h26.71 for tipranavir, purchased in com-munity pharmacies.

Antiretroviral-Related Costs

The estimates of the annual antiretroviral-related costs incurred by patients in each of themodelled scenarios (with and without DRV/r)were derived partly from an analysis of thehealthcare resource consumption and relatedcosts of highly treatment-experienced HIV-infected patients included in the FHDH database.In the cost-of-care study, described elsewhere inthis supplement,[29] estimates of the annual anti-retroviral-related costs were calculated for pa-tients at different levels of HIV-induced immunedeficiency, corresponding to the CD4 cell countstrata considered by the model (table III).

These costs were divided into PI-related andnon-PI-related costs. The model then calculatedthe estimated weighted, average, annual PI-relatedcosts of patients during each year of the model inboth scenarios, based on the number of patientsestimated to be using DRV/r, TPV/r and otherPIs in the target population at that time. For thelatter group, the average annual PI-related cost,derived from the cost-of-care study was used. Forpatients on TPV/r and DRV/r, the annual per-patient cost was calculated bymultiplying the dailyacquisition cost of TPV/r and DRV/r by 365 days,

Table III. Estimated annual HIV-related healthcare expenditure per patient as a function of CD4 cell count range in the baseline case (current

HIV treatment strategy)

CD4 cell count range (cells/mm3)

A B C D E F All

(£50) (51–100) (101–200) (201–350) (351–500) (>500)

Average yearly antiretroviral expenditure/patient (Euros)

PIs 5380 4885 4539 4010 3632 3277 3973

Other antiretroviral drugs 7692 7518 7581 7463 7405 7368 7691

Total antiretroviral drugs 13 072 12403 12120 11473 11037 10 645 11 664

Other expenditure (Euros)

Other HIV-related drugs 8970 6710 3255 2050 1312 734 3145

Hospitalizations (inpatient stays) 11 536 7682 3234 1878 1109 678 3418

Day hospital 814 1006 805 651 623 391 689

Hospital consultations (collected in the FHDH

database)

34 40 37 36 35 24 35

Total non-antiretroviral costs 21 354 15438 7332 4615 3080 1826 7286

Data are taken from the FHDH database.[30]

FHDH =French Hospital Database in HIV; PIs =protease inhibitors.



obtained from drug lists available for hospitaland community pharmacies.[31] The averagetotal, annual antiretroviral-related cost of care perpatient in each year of the model in each scenariowas then calculated by adding the weightedannual PI-related cost per patient to the non-PI-related cost. The model assumed that the in-troduction of darunavir would not impact thedistribution of the utilization (and thus cost) ofother antiretroviral classes in the modelled targetpopulation.

Non-Antiretroviral-Related Costs

The average, annual non-antiretroviral-related costs incurred by patients in each of theCD4 cell count strata considered by the modelwere also derived from the cost-of-care analysis,mentioned above, that was performed on theFHDH database. As shown (table III), the non-antiretroviral-related cost-of-care drivers includedin this cost study were non-antiretroviral-relatedmedications, inpatient and day hospitalizationsand those consultations ‘intra muros’ that wererecorded in the FHDH database.[30]

Budget Impact Calculation

Finally, the budget impact was estimated asthe difference in the total cost (antiretroviral andnon-antiretroviral-related) of the scenarios withand without darunavir in each year of the modeland over the entire modelled time horizon.

Sensitivity Analyses

One-way sensitivity analyses were performedin order to determine the sensitivity of the esti-mated budget impact to varying different inputsof the model. The different inputs were the size ofthe modelled study population and its growthrate over time, forecasted market shares ofdarunavir, uptake of tipranavir in both modelledscenarios, baseline CD4 cell count distribution ofthe modelled study population (based on datafrom patient subgroups with a varying degree ofprevious PI failure included in the FHDH studysample) and the immunological efficacy of dar-unavir-basedHAART as observed in the POWER1 and 2 trials (600/100mg bid patient subgroups).In all these analyses, the influence of these mod-ified inputs on the estimated budget impact dif-ference in total expenditures between the twomodelled scenarios (with and without darunavir)was evaluated over the entire 3-year time horizonof the model. The external validity of the esti-mates was evaluated with respect to recent pub-lished information on healthcare expenditureassociated with the management of HIV infectionin France.[32,33]

Results

CD4 Distribution in Each of the ModelledScenarios

Table IV shows the anticipated proportionaldistributions of patients in the different CD4 cell

Table IV. Predicted CD4 cell count distributions in years 1, 2 and 3 for scenarios with and without darunavir becoming available for the

treatment of the target population in year 1

CD4 cell count range (cells/mm3) Darunavir not available Darunavir available from year 1

Year 1 (%) Year 2 (%) Year 3 (%) Year 1 (%) Year 2 (%) Year 3 (%)

0–50 10.8 10.2 9.2 9.4 7.5 5.6

51–100 8.4 8.1 7.6 7.6 6.4 5.2

101–200 19.1 18.9 18.3 18.3 16.6 15.1

201–350 28.1 28.2 28.2 28.0 27.7 27.5

351–500 19.3 19.6 20.2 20.3 21.7 23.1

>500 14.3 15.0 16.5 16.4 20.0 23.5

Values shown derive from the transition state model that was constructed using efficacy data predictions and the forecasted evolution of the

patient shares of the different protease inhibitor options.

190 Colin et al.


count ranges for the three modelled years in eachof the two scenarios considered by the model,based on the efficacy data and the forecastedevolution of the patient shares of the different PIoptions. In the scenario in which darunavir be-came available from year 1 for the treatment ofthe target population, the proportion of patientsin the lower, more costly CD4 cell count strata(£100 cells per mm3) was consistently lowerthan in the scenario without darunavir in eachyear of the model (17.0% vs 19.2%, 13.9% vs18.3% and 10.8% vs 16.8% in years 1, 2 and 3,respectively).

Total Healthcare Costs in Each ModelledScenario

The anticipated evolution of the annual anti-retroviral-related, non-antiretroviral-related andtotal costs associated with the treatment andclinical management of the target population inyears 1, 2, 3 and over the entire time horizon of

the model in both modelled scenarios in the base-case analysis is detailed in table V.

Compared with the scenario in which dar-unavir would not be available for the treatmentof the target population, the overall healthcarebudget for these patients is expected to decreaseas incremental antiretroviral-related costs arefully offset by savings in non-antiretroviral-related costs, even during the first year. Over theentire 3-year period, the net increase in antire-troviral drug costs (h+5.6 million) is expected tobe fully compensated by savings in hospitaliza-tion costs (h-9.7 million) and in the expendituresfor HIV-related medications other than antire-troviral drugs (h-7.3 million), leading to a netbenefit of h11.4 million (or a relative saving of-2.9% of the total budget for the management ofthe modelled population) compared with thescenario in which darunavir would not be avail-able in the base-case analysis.

The results of the one-way sensitivity analysesperformed as part of this study are reported in

Table V. Total healthcare expenditure (in Euros) associated with following the current treatment strategy and following the darunavir

treatment strategy over the lifetime of the model

Base-case Year 1 Year 2 Year 3 3-Year period

Darunavir not available (1)

Antiretroviral costs 69 749 645 61.5% 75020 612 62.2% 81 853 708 63.2% 91 212618 65.3% 248086 938 63.6%

Other HIV-related

drugs

18 806 546 16.6% 19674 933 16.3% 20 550 904 15.9% 20 952827 15.0% 61178 665 15.7%

Hospitalizations

and follow-up costs

24 766 533 21.9% 25909 118 21.5% 27 051 580 20.9% 27 565984 19.7% 80526 682 20.7%

Total costs 13 322 724 100.0% 120604 663 100.0% 129 456 192 100.0% 139 731429 100.0% 389792 285 100.0%

Darunavir available (2)

Antiretroviral costs 69 749 645 61.5% 76486 786 64.0% 84 349 510 67.0% 92 890739 69.9% 253727 035 67.1%

Other HIV-related

drugs

18 806 546 16.6% 18579 948 15.5% 17 976 725 14.3% 17 263692 13.0% 53820 365 14.2%

Hospitalizations

and follow-up costs

24 766 533 21.9% 24453 889 20.5% 23 654 983 18.8% 22 705269 17.1% 70814 141 18.7%

Total costs 113 322 724 100.0% 119520 623 100.0% 125 981 218 100.0% 132 859700 100.0% 378361 542 100.0%

Budget impact (1)–(2)

Input variable Difference in expenditure (Euros)

Antiretroviral costs 1 466 173.55 2 495 801.63 1 678121.47 5 640 096.64

Other HIV-related

drugs

-1094 985.29 -2 574 179.03 -3 689135.22 -7358 299.54

Hospitalizations

and follow-up costs

-1455 228.53 -3 396 596.75 -4 860715.06 -9712 540.33

Total costs -1084 040.27 -3 474 974.15 -6 871728.81 -11430 743.23



table VI. The most important parameter was thebaseline distribution of CD4 cell count in thetarget population. A large variation in the bud-getary impact over the entire 3-year time horizonwas observed when the anticipated market shareof tipranavir in a scenario with darunavir wasprogressively increased over the time horizon ofthe model and when the clinical efficacy of DRV/r-based HAART was varied, considering the netincrease in the CD4 cell count from baseline ofboth POWER 1 and 2 clinical trials separately.[34]

Nevertheless, in all sensitivity analyses, the use ofdarunavir resulted in savings of the estimated,total direct costs associated with the managementof the modelled population, comparative to ascenario in which darunavir would not have

been available during the time horizon of themodel.

Discussion

This budget impact model analysis suggeststhat the introduction of DRV/r in the manage-ment of highly treatment-experienced, HIV-infected adults seeking care in France is expectedto generate a small increase in the antiretroviral-related costs of care associated with this popula-tion over the period 2007–2009. This increasein ART costs is, however, expected to be fullyoffset by a reduction in hospitalization-relatedexpenditures and acquisition costs for othermedications used in this population, as a result of

Table VI. Sensitivity analysis of the model

Parameters Base-case

analysis

Test case Base-case

analysis

Test case Percentage

Target population size at model entry 5980 4000 -11430 743 -7644 432 +33

8000 -15290 324 -34

Rate of growth of target population 6.6% 0% -11430 743 -9746 507 +15

10% -12368 408 -8

Market share of darunavir (no change

in tipranavir market share)

Year 1/year2/year 3:20%/45%/70%

Relative decrease:

-30% for each year

-11430 743 -9288 152 +19

Relative decrease:

-15% for each year

-10355 244 +9

Relative increase:

+15% for each year

-12579 474 -10

Relative increase:

+30% for each year

-13682 694 -20

Market share of tipranavir in scenario

with darunavir entering the market

(no change in darunavir market share)

5% (years 1, 2

and 3)

10% (year 1)

15% (year 2)

25% (year 3)

-11430 743 -7095 740 +38

Baseline CD4 cell count distribution

according to number of previous PI

failures

All patients in

study sample

1 PI failure -11430 743 -6568 535 +43

2 PI failures -9906 517 13

‡3 PI failures -15722 559 -38

DRV/r clinical efficacy expressed as

change in CD4 cell count (mean –SD) at24 and 48 weeks

92.4 – 113.6102.0 – 126.9(pooled data from

POWER trials

600/100mg bid

subgroups)[34]

118 (–136.2)130 (–146.0)(POWER 1 data

only)[35]

-11430 743 -15856 793 -39

67.3 (–79.1)77.0 (–100.6)(POWER 2 data

only)[36]

-7614 377 +33

POWER 1 and 2 clinical trials analyzed separately.

DRV/r =darunavir/ritonavir; PI =protease inhibitor; POWER =Performance Of TMC114/r When evaluated in treatment-Experienced patients

with protease inhibitor Resistance; SD = standard deviation.

192 Colin et al.


the significant immune reconstitution of patientstreated with DRV/r-based HAART. CD4 cellcounts would thus increase to levels that pose alesser risk for developing costly AIDS-definingopportunistic events and cancers and other sig-nificant clinical events such as liver disease,non-AIDS-defining cancers, hepatitis and cardio-vascular pathologies.[22,37] Based on these results,the use of darunavir-based HAART regimens inhighly treatment-experienced, HIV-infected adultsappears to be economically attractive because itoffers the potential to reduce the total expendi-tures for these patients in France, at a time whenHIV budgets, like other components of the Frenchhealthcare budget, are coming under increasedpressure.

A number of sensitivity analyses was perform-ed in order to verify whether the positive budgetimpact associated with the use of darunavir wouldbe maintained if various input variables used inthe model were changed. In all of these cases, thepositive impact remained and no scenario wasidentified in which the introduction of darunavirwould incur incremental total expenditures to thehealthcare system. As expected, the most positivebudget impact was observed in those patients whohad failed multiple PI-based treatments and werein a more advanced stage of HIV-induced im-mune deficiency. The anticipated market share oftipranavir may also be considered as an importantparameter, as a higher uptake of TPV(/r) in thedarunavir scenario would lead to lower savings asa result of the higher price and lower efficacycompared with darunavir.

In common with other economic analyses, ourstudy has a number of limitations. First, as TPV/rwas not included as a treatment option in thePOWER 1 and 2 trials because it was not com-mercially available at the time, we used estimatesof the CD4 cell count increase from baseline as-sociated with DRV/r and TPV/r-based HAARTregimens that were derived from two separateclinical trials, the pooled POWER 1 and 2 (forDRV/r) and RESIST 1 and 2 trials (for TPV/r) tomodel the efficacy of these regimens in the targetpopulation. Whereas direct comparative clinicalevidence would have provided a more robustbasis for the modelling, our approach was consid-

ered adequate because the RESIST and POWERtrials had a similar study design and eligibilitycriteria and the baseline characteristics of theenrolled patient populations have been shown tobe broadly comparable.[18,23,27]

Second, we calculated the predicted shift inCD4 cell count distribution associatedwith the useof DRV/r and TPV/r-based HAART by system-atically applying the average increase in CD4 cellcount from baseline observed in the POWER andRESIST trials across all baseline levels of CD4 cellcount. The efficacy of ARTmay, however, vary asa function of the starting CD4 cell count, as sug-gested by earlier studies and the higher mean in-crease in CD4 cell count observed in the POWER1 trial that recruited patients with less advancedimmune deficiency than in POWER 2.[34-36,38]

Third, although healthcare costs have beenshown to be higher in patients with lower CD4cell counts, no French study has formally dem-onstrated that increasing CD4 cell counts willreduce resource consumption. Nonetheless, thestudy of Flori and le Vaillant[39] on the FHDHdatabase showed that the total annual cost ofcare for HIV patients decreased over time whenHAART therapy was initiated systematically,leading to a shift in the CD4 cell count distribu-tion towards higher values. In this study, theassociated increase in drug acquisition costs forHAART was outweighed by the decrease inhospital costs (complete stays). This observationis consistent with the outcome of our modellingstudy with darunavir.

The estimates of the annual average total ex-penditures per patient that we used in our modelare higher than previously published estimates,but these differences can be explained by the fo-cus on a more advanced and highly treated HIVpopulation. In this model, certain classes of costassociated with the management of HIV patients,such as community care consultations and trans-portation, were not included because of lack ofinformation in the FHDH database. The relation-ship between the level of expenditure on thesecost drivers and the CD4 cell count is poorlycharacterized. Indeed, Detournay et al.[40] haveshown that only hospital costs were statisticallyrelated to the severity of the disease expressed in



CD4 cell counts with a threefold difference be-tween patients with less than 200 cells/mm3 andthose with more than 500 cells/mm3.[40] Otherauthors have, however, shown a relationship be-tween other types of expenditure withHIV diseaseseverity.[41] If the level of community care ex-penditures and transportation costs is unrelatedto CD4 cell count, then taking these into accountwould dilute the budget impact of darunavir es-timated in our model. As hospital costs and thetypes of medication costs considered in ourmodel have been shown to represent 83–93% ofthe total HIV cost for the whole population inFrance, the impact of the cost drivers that werenot included in the model on the conclusion ofour analysis may be limited.[32,33]

Finally, when we calculated the budget impactof darunavir, we assumed that its introductionwould not impact the distribution of utilizationof other existing antiretroviral classes. As a num-ber of novel antiretroviral agents, some belongingto new drug classes, have recently been approvedbased on the results of clinical trials in highlytreatment-experienced patient populations, treat-ment practices and related expenditures mayevolve.[15-18,42] Our model does not take this intoconsideration. In most of the trials of the newagents, however, they were added to optimizedHAART regimens containing ritonavir-boostedPIs (including DRV/r 600/100mg bid), reversetranscriptase inhibitors and, if needed enfuvir-tide, rather than replacing any of these components.Therefore, our assumptions may remain valid,especially given the time horizon of the model(2007–2009).

Despite the limitations described above, ourstudy offers valuable information. Analyses ofthe economic impact of antiretroviral medica-tions typically involve cost-effectiveness model-ling using quality-adjusted life-years (QALYs) orother disease-specific endpoints.[43] Managersof healthcare systems nevertheless also need in-formation on the potential impact of the intro-duction of new medications on their budget inorder to plan resources and expenditures. To date,only a limited number of formal analyses of theimpact of therapeutic and other interventions onthe HIV care budget at a population level have

been published, and none of them have focusedon the impact of one particular medication.[44,45]

Elsewhere in this supplement, Hill et al. haveevaluated the impact of the use of DRV/r600/100mg bid, compared with currently avail-able PIs, on the overall (antiretroviral and non-antiretroviral-related) costs of care in highlypre-treated, POWER-like patients during thefirst patient-year of therapy in the USA, the UK,Belgium, Italy and Sweden.[46] In their analysis,the authors compared a scenario in which an en-tire cohort of patients either takes darunavir-based HAART or a regimen containing other PIs(excluding tipranavir). The impact of a medica-tion on a payer’s budget will, however, usuallyvary over time with its degree of utilization inthe target population. Therefore, we decided toadopt a different approach and allowed the up-take (and thus the budget impact) of darunavir bythe target population to vary over a 3-year timehorizon.

It should be stressed that the conclusions ofour analysis only apply to the highly pre-treatedtarget population included in this model. In thephase III TITAN trial (TMC114-C214; TMC114/rIn Treatment-experienced pAtients Naive to lopi-navir), which enrolled less treatment-experiencedpatients than the POWER trials, the superiorvirological response observed after 48 weeks oftherapy with DRV/r 600/100mg bid, comparedwith lopinavir/ritonavir (both given with an OBR),was not associated with a higher CD4 cell countincrease from baseline for DRV/r.[47] This impliesthat the favourable impact of DRV/r 600/100mgbid on the non-antiretroviral-related costs of careobserved in our analysis may not be seen in lesstreatment-experienced patients, at least not in thefirst year. In TITAN, DRV/r-based HAART wasassociated with a lower rate of premature treat-ment discontinuation during the first 48 weeks oftherapy (21% vs 29%), which appeared to beprimarily driven by a higher rate of virologicalfailure observed in the lopinavir/ritonavir arm(11% vs 1%) and an increased risk of the devel-opment of additional resistance-associated mu-tations (RAMs) and loss of susceptibility to PIsand/or nucleoside reverse transcriptase inhibitors(NRTIs).[47] As current HIV treatment guidelines

194 Colin et al.


recommend that, in the case of treatment failure,a newHAART regimen with at least two and pre-ferably three active agents should be constructedto ensure the long-term success of the therapyand the presence of (cross)resistance to PIs andNRTIs can significantly compromise this,[10,11,48]

DRV/r 600/100mg bid may therefore potentiallygenerate savings in ART costs in earlier lines oftherapy in France by preserving future PI andNRTI treatment options and reducing the needfor more complex and expensive regimens such asthose containing the injectable antiretroviraldrug enfuvirtide or the recently approved novel,oral antiretroviral agents maraviroc and ralte-gravir.[15-17,49,50] This, however, remains to befurther investigated.

The phase III development programme ofDRV/r is currently also evaluating a once-dailydose of 800mg with 100mg ritonavir in treat-ment-experienced patients who harbour no dar-unavir RAMs. Such patients accounted for 82%of the overall population included in the TITANtrial.[47,51] This lower DRV/r dose has shownimpressive efficacy results and a favourable tol-erability profile in antiretroviral-naive patientsand in the subgroup of POWER patients with nodarunavir RAMs in whom it generated a viro-logical response that was comparable to thatobserved with 600/100mg bid (62% vs 67%; dif-ference 5; 95% confidence interval -30 to 23), withboth regimens being superior to control (11%;p< 0.0001).[52,53] Moreover, in addition to con-venient once-daily dosing, it also has a lower pillburden than the currently approved 600/100mgbid dose of DRV/r (three vs six pills per day, in-cluding ritonavir). As lowering the dosing fre-quency – in particular once-daily dosing – and pillburden have been shown to impact adherence toART positively and thereby improve its long-term outcome,[54-56] the future approved use ofthis lower, once-daily dose of DRV/r in the ma-jority of treatment-experienced, HIV/AIDS pa-tients in routine clinical practice may potentiallylead to a (further) improvement of the impact ofDRV/r-based HAART on the overall costs ofcare of the treatment-experienced population inFrance. Here again, further research is, however,needed to confirm this hypothesis.

Conclusions

In summary, in comparison to a scenario with-out darunavir, the use of DRV/r 600/100mg bidin combination with other antiretroviral agents,in highly pre-treated, HIV-infected adult patientswho have failed one or more PI-containingHAART regimen in France during 2007–2009,is anticipated to increase the antiretroviral drugexpenditures for these patients as a result of thereplacement of older and cheaper PIs by dar-unavir. This increase, however, would be offset bya reduction in hospital costs and other HIV-relatedmedication costs, providing a small decreasein the total expenditures for this population.

The use of DRV/r 600/100mg bid in this pa-tient population, which represents one of themost costly HIV subpopulations of patients tomanage, is therefore expected to cause no in-crease in the budget of the French Sickness Fund.This conclusion was confirmed by the results ofseveral sensitivity analyses. Further research isneeded to estimate the budget impact of the useof DRV/r in less treatment-experienced, HIV-infected individuals in France.

Acknowledgements

The authors wish to acknowledge Anita Brogan (RTIHealth Solutions, Research Triangle Park, North Carolina,USA) for her involvement in the development and program-ming of the model used in this analysis. Their gratitude alsogoes to Jean-Marie Lang (Clinique Medicale A, StasbourgUniversity Hospital, Strasbourg, France) for his assistance inthe analysis of the French Hospital Database in HIV (FHDH)cost-of-care data. Catherine Elliott (medical writer, Gardiner-Caldwell Communications, Macclesfield, UK) provided edi-torial support and assisted the authors in responding to peerreview comments. Finally, the authors wish to thank the in-vestigators involved and, last but not least, the patients in-cluded in the FHDH database as well as the clinical trials thatwere used as sources for this analysis. This project was fi-nancially supported by Janssen-Cilag SAS, France.

XC and AL declare that this study was funded by Johnson& Johnson (Janssen Cilag). ES is an employee of Johnson &Johnson Pharmaceutical Services, Beerse, Belgium, and ownsstock options and shares in this company. JM has receivedgrant support from Janssen Cilag, the manufacturer of darunavir,to assist with the preparation of this manuscript. JM was notrestricted by Janssen Cilag in her interpretation of the individ-ual papers on which the review was based. PG is an employeeof Janssen Cilag and declares that this project was financiallysupported by Janssen Cilag.



References1. Palella FJ, Delaney KM, Moorman AC, et al. Declining

morbidity and mortality among patients with advancedhuman immunodeficiency virus infection. N Engl J Med1998; 338: 853-60

2. Walensky RP, Paltiel D, Losina E, et al. The survival bene-fits of AIDS treatment in the United States. J Infect Dis2006; 194: 11-9

3. Pomerantz RJ, Horn DL. Twenty years of therapy for HIV-1infection. Nat Med 2003; 9: 867-73

4. Schackman BR, Gebo KA, Walensky RP, et al. The lifetimecost of current human immunodeficiency virus care in theUnited States. Med Care 2006; 44: 990-7

5. Mouton Y, Alfandari S, Valette M, et al. Impact of proteaseinhibitors on AIDS-defining events and hospitalizations in10 French AIDS reference centres. Federation Nationaldes Centres de Lutte contre le SIDA.AIDS 1997; 11: F101-5

6. Mocroft A, Ledergerber B, Katlama C, et al. Decline in theAIDS and death rates in the EuroSIDA Study: an obser-vational study. Lancet 2003; 362: 22-9

7. Gardner LI, Klein RS, Szczech LA, et al. Rates and riskfactors for condition-specific hospitalizations in HIV-infected and uninfected women. J Acquir Immune DeficSyndr 2003; 34: 320-30

8. Torti C, Casaria S, Palvarini L, et al. Modifications of healthresource-use in Italy after the introduction of highly active anti-retroviral therapy (HAART) for human immunodeficiencyvirus (HIV) infection. Pharmaco-economic implications in apopulation-based setting. Health Policy 2003; 65: 261-7

9. Gallant JE. Approach to the treatment-experienced patient.Infect Dis Clin North Am 2007; 21: 85-102

10. Department of Health and Human Services (DHHS) Panelon antiretroviral guidelines for adults and adolescents. Guide-lines for the use of antiretroviral agents in HIV-infectedadults and adolescents; 2009 Dec 1 [online]. Available fromURL: http://aidsinfo.nih.gov/ContentFiles/AdultandAdolescentGL.pdf [Accessed 2009 Dec 10]

11. BHIVA Writing Committee. British HIV Association(BHIVA) draft guidelines for the treatment of HIV-infectedadults with antiretroviral therapy (2006). HIVMed 2006; 7:487-503

12. European AIDS Clinical Society (EACS). 2007 Guidelinesfor the clinical management and treatment of HIV-infectedadults in Europe; 2009 Nov [online]. Available from URL:http://www.europeanaidsclinicalsociety.org/guidelines.asp[Accessed 2009 Dec 10]

13. Food and Drug Administration (FDA): Press Release: FDAapproves darunavir; 2006 Jun 23 [online]. Available fromURL:www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/2006/ucm108676.htm [Accessed 2009 Dec 10]

14. European Commission Community register of medicinalproducts for human use. PREZISTA; [online]. Availablefrom URL: http://ec.europa.eu/enterprise/sectors/pharmaceuticals/documents/community-register/html/h380.htm[Accessed 2009 Dec 10]

15. Food and Drug Administration (FDA): press release. FDAapproves novel antiretroviral drug; 2007 Aug 6 [online].Available from URL: http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/2007/ucm108960.htm[Accessed 2009 Dec 10]

16. Food and Drug Administration (FDA): press release. FDAapproves new antiretroviral drug; 2007 Oct 16 [online].Available from URL: http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/2007/ucm109008.htm[Accessed 2009 Dec 10]

17. Committee forMedicinal Products forHumanUse (CHMP) ofthe European Medicines Evaluation Agency (EMEA). Sum-mary of positive opinion for raltegravir. 2007 Nov 27 [online].Available from URL: http://www.emea.europa.eu/pdfs/human/press/pr/52210907en.pdf [Accessed 2009 Dec 10]

18. Clotet B, Bellos N, Molina JM, et al. Efficacy and safety ofdarunavir–ritonavir at week 48 in treatment-experiencedpatients with HIV-1 infection in POWER 1 and 2: a pooledsubgroup analysis of data from two randomised trials.Lancet 2007; 369: 1169-78

19. Moeremans K, Annemans L, Lothgren M, et al. Cost effec-tiveness of darunavir/ritonavir 600/100mg bid in proteaseinhibitor-experienced, HIV-1-infected adults in Belgium,Italy, Sweden and the UK. PharmacoEconomics 2010; 28Suppl. 1: 107-28

20. Mauskopf J, Brogan A, Martin S, et al. Cost effectiveness ofdarunavir/ritonavir in highly treatment-experienced, HIV-1-infected adults in the USA. PharmacoEconomics 2010;28 Suppl. 1: 83-105

21. BroganA,Mauskopf J, Talbird SE, et al. US cost effectiveness ofdarunavir/ritonavir 600/100mg bid in treatment-experienced,HIV-infected adults with evidence of protease inhibitorresistance included in the TITAN trial. PharmacoEco-nomics 2010; 28 Suppl. 1: 129-46

22. Holkmann Olsen C, Gatell J, Ledergerber B, et al. Risk ofAIDS and death at given HIV-RNA and CD4 cell counts,in relation to specific antiretroviral drugs in the regimen.AIDS 2005; 19: 319-30

23. Hicks C, Cahn P, Cooper DA, et al. Durable efficacy oftipranavir–ritonavir in combination with an optimisedbackground regimen of antiretroviral drugs for treatment-experienced HIV-1-infected patients at 48 weeks in theRandomized Evaluation of Strategic Intervention in multi-drug resistant patients with Tipranavir (RESIST) studies:an analysis of combined data from two randomised open-label trials. Lancet 2006; 368: 466-75

24. Comite Economique des Produits de Sante 2008 ArticleL162-17-3; [online]. Available from URL: http://www.sante.gouv.fr/ceps [Accessed 2009 Dec 10]

25. Base de donnees Hospitaliere Francaise sur l’infection a VIH –ANRS CO4 FHDH. Inserm transfert SA, Paris, France, 2007

26. La lutte contre l’epidemie de VIH-SIDA en France – Com-munication du ministre de la Sante et des Solidaritesrelative a la lutte contre l’epidemie de VIH SIDA en France– Conseil des ministres du 28 novembre 2006; [online].Available fromURL: http://www.archives.premier-ministre.gouv.fr/villepin/acteurs/gouvernement/conseils_ministres_35/conseil_ministres_28_novembre_871/lutte_epidemie_vih_sida_57330.html [Accessed 1009 Dec 10]

27. Hill A, Moyle G. Relative antiviral efficacy of ritonavir-boosted darunavir and ritonavir-boosted tipranavir vs.control protease inhibitor in the POWER and RESISTtrials. HIV Med 2007; 8: 259-64

28. Boehringer Ingelheim. AptivusTM (tipranavir) PrescribingInformation; revised 2008 June [online]. Available fromURL: http://bidocs.boehringer-ingelheim.com/BIWebAc

196 Colin et al.


cess/ViewServlet.ser?docBase=renetnt&folderPath=/Prescribing+Information/PIs/Aptivus/10003515+US+01.pdf[Accessed 2009 Dec 10]

29. Colin X, Lafuma A, Costagliola D, et al. The cost ofmanaging HIV infection in highly treatment-experienced,HIV-infected adults in France. PharmacoEconomics 2010;28 Suppl. 1: 59-68

30. GERS: Groupement pour l’Elaboration et la Realisation deStatistiques. Nos missions [online]. Available from URL:http://www.gie-gers.fr/groupement/qui_sommes_nous.php3[Accessed 2009 Dec 10]

31. Base de donnees Hospitaliere Francaise sur l’infection a VIH –ANRSCO4FHDH. Inserm transfert SA, Paris, France, 2006[online]. Available from URL: www.invs.sante.fr/beh/2007/46_47/index.htm [Accessed 2009 Dec 10]

32. Weill A, Chinaud F, Vallier N, et al. Frequence et couts destrentes affections de longue duree de la region Midi-Pyrenees, annee 2003. Revue Medicale de l’AssuranceMaladie 2005; 4: 273-87

33. Vallier N,Weill A, Salanave B, et al. Cost of thirty long-termdiseases for beneficiaries of the French general health in-surance scheme in 2004. Pratique et Organisation des Soins2006; 37: 267-83

34. Tibotec Pharmaceuticals. Clinical data of POWER clinicaltrial for darunavir/r. Data on file, 2007 Dec

35. KatlamaC, EspositoR, Gatell JM, et al. Efficacy and safety ofTMC114/ritonavir in treatment-experienced HIV patients:24-week results of POWER 1. AIDS 2007; 21: 395-402

36. Haubrich R, Berger D, Chiliade P, et al. Week 24 efficacyand safety of TMC114/ritonavir in treatment-experiencedHIV patients. AIDS 2007; 21: F11-18

37. Weber R, Friis-Moller N, Sabin C, et al. HIV and non-HIV-related deaths and their relationship to immuno-deficiency: The D:A:D Study [abstract 595]. 12thConference on Retroviruses and Opportunistic Infections;2005 Feb 22–25; Boston, MA, USA

38. Hirsch HH, Kaufmann G, Sendi P, et al. Immune recon-stitution in HIV-infected patients. Clin Infect Dis 2004; 38:1159-66

39. Flori YA, le Vaillant M. Use and cost of antiretrovirals inFrance 1995–2000: an analysis based on the Medical Dos-sier on Human Immunodeficiency (release 2) database.PharmacoEconomics 2004; 22: 1061-70

40. Detournay B, Voinet C, Bez J, et al. Consommation medicaledes patients infectes par le VIH suivis a l’Hopital Purpan(Toulouse). Journal D’Economie Medicale 2003; 21: 367-76

41. Gebo KA, Chaisson RE, Folkemer JG, et al. Costs of HIVmedical care in the era of highly active antiretroviral ther-apy. AIDS 1999; 13: 963-9

42. U.S. Food and Drug Administration (FDA). 2008 [online].Available from URL: http://www.fda.gov/oashi/aids/status.html [Accessed 2009 Dec 10]

43. Ruof J, Dusek A, DeSpirito M, et al. Cost-efficacy com-parison among three antiretroviral regimens in HIV-1 in-fected, treatment-experienced patients. Clin Drug Invest2007; 27: 469-79

44. Schackman BR, Freedberg KA, Goldie SJ, et al. Budgetimpact of Medicaid Section 1115 demonstrations for earlyHIV treatment. Health Care Financ Rev 2005; 26: 67-80

45. Sansom SL, Anthony MN, Garland WH, et al. The costs ofHIV antiretroviral therapy adherence programs and im-pact on health care utilization. AIDS Patient Care STDS2008; 22: 131-8

46. Hill A, Gebo K, Hemmett L, et al. Predicting direct costs ofHIV care during the first year of darunavir-based highlyactive antiretroviral therapy using CD4 cell counts: evidencefrom POWER. PharmacoEconomics 2010; 28 Suppl. 1:169-81

47. Madruga JV, Berger D, McMurchie M, et al. Efficacy andsafety of darunavir–ritonavir compared with that of lopi-navir–ritonavir at 48 weeks in treatment-experienced, HIV-infected patients in TITAN: a randomised controlled phaseIII trial. Lancet 2007; 370: 49-58

48. International AIDS Society –USA Panel. Treatment for adultHIV infection: 2006 recommendations of the InternationalAIDS Society – USA Panel. JAMA 2006; 296: 827-43

49. Horne R, Clotet B, Cohen C, et al. Treatment-experiencedpatient perceptions of self-injectable therapy [abstractPE7.3/25]. 10th European AIDS Clinical Society Con-ference; 2005 Nov 11–14; Dublin, Ireland

50. Committee forMedicinal Products for HumanUse (CHMP)of the European Medicines Evaluation Agency (EMEA).Summary of positive opinion for maraviroc. 2007 [online].Available from URL: http://www.emea.europa.eu/humandocs/Humans/EPAR/celsentri/celsentri.htm [Acces-sed 2009 Dec 10]

51. Berger DS, Northland R, Scribner A, et al. Effect of baselinefactors on virological response to darunavir/r and lopinavir/rat week 48 in TITAN [abstract P7.3/27]. 11th EuropeanAIDS Conference; 2007 Oct 24–27; Madrid, Spain

52. Ortiz R, Dejesus E, Khanlou H, et al. Efficacy and safety ofonce-daily darunavir/ritonavir versus lopinavir/ritonavir intreatment-naıve HIV-1-infected patients at week 48. AIDS2008; 22: 1389-97

53. Tibotec Pharmaceuticals. Data on file, 2007

54. Ammassari A, Trotta MP, Murri R, et al. Correlates andpredictors of adherence to highly active antiretroviraltherapy: overview of published literature. J Acquir Im-mune Defic Syndr 2002; 31 Suppl. 3: S123-7

55. Maggiolo F, Ripamonti D, Airoldi M, et al. Resistance costsand future drug options of antiretroviral therapies: analysisof the role of NRTIs, NNRTIs, and PIs in a large clinicalcohort. HIV Clin Trials 2007; 8: 9-18

56. Negredo E, Bonjoch A, Clotet B. Benefits and concerns ofsimplification strategies in HIV-infected patients. J Anti-microb Chemother 2006; 58: 235-42

Correspondence: Dr Xavier Colin, Cemka-Eval, 43Boulevard du Marechal Joffre, 92340 Bourg-la-Reine,France.E-mail: [email protected]



modelling the budget impact of darunavir in the treatment of highly treatment-experienced,...

Documents