us cost effectiveness of darunavir/ritonavir 600/100mg bid in treatment-experienced, hiv-infected...

US Cost Effectiveness of Darunavir/Ritonavir 600/100mg bid in Treatment-Experienced, HIV-Infected Adults withEvidence of Protease Inhibitor ResistanceIncluded in the TITAN TrialAnita Brogan,1 Josephine Mauskopf,1 Sandra E. Talbird1 and Erik Smets2

1 RTI Health Solutions, Research Triangle Park, North Carolina, USA

2 Johnson & Johnson Pharmaceutical Services LLC, Beerse, Belgium

Abstract Introduction: The phase III TITAN trial evaluated the use of darunavir with

low-dose ritonavir (DRV/r) 600/100mg twice daily (bid) compared with lo-

pinavir with low-dose ritonavir (LPV/r) in treatment-experienced, lopinavir-

naive patients. This study estimates the cost effectiveness of DRV/r from a

US societal perspective when compared with LPV/r in treatment-experienced

patients with a profile similar to those TITAN patients who had one or more

International AIDS Society – USA (IAS–USA) primary protease inhibitor

(PI) resistance-associated mutations (RAMs) at baseline. This population

had less advanced HIV disease and a broader range of previous PI exposure/failure (0–‡2 PIs) at enrolment than those in the darunavir phase IIb POWER

trials.

Methods: An existing Markov model containing six health states defined

by CD4 cell count range (>500, 351–500, 201–350, 101–200, 51–100 and

0–50 cells/mm3) and an absorbing state of death was adapted. Baseline de-

mographics, CD4 cell count distribution and antiretroviral drug usage, vi-

rological response (at week 24), and immunological response estimates and

matching transition probabilities were based on data collected directly from

the one or more IAS–USA PI mutation subpopulation during the first

48 weeks of the TITAN trial, as well as from published literature. Patients

were assumed to switch to a regimen containing tipranavir plus an optimized

background regimen after treatment failure. For each CD4 cell count range

or health state, the utility values, HIV and non-HIV-related mortality rates,

and non-antiretroviral-related cost of HIV care estimates were derived from

published literature. Unit costs were derived from official local sources.

A lifetime horizon was taken in the base-case analysis.

ORIGINAL RESEARCH ARTICLEPharmacoeconomics 2010; 28 Suppl. 1: 129-146

1170-7690/10/0001-0129/$49.95/0

ª 2010 Adis Data Information BV. All rights reserved.

Results: The base-case analysis predicted discounted quality-adjusted survi-

val gains of 0.493 quality-adjusted life-years (QALYs) for DRV/r compared

with LPV/r, resulting in an incremental cost-effectiveness ratio (ICER) of

US$23 057 per QALY gained over a lifetime horizon. Probabilistic sensitivity

analysis indicated a 0.754 probability of an ICER below the threshold of

US$50 000 per QALY gained. DRV/r remained cost effective over all para-

meter ranges tested in extensive one-way sensitivity analyses and variability

analyses, which examined the impact of input parameter uncertainty and

changes in model assumptions and treatment patterns, respectively. Shorten-

ing the model time horizon had the largest impact on the ICER, reducing it

most notably to US$4919 with a 10-year time horizon.

Conclusion: From a US societal perspective and based on an analysis of the

patients with primary IAS–USAPIRAMs enrolled in the darunavir phase III

TITAN trial, a highly active antiretroviral therapy (HAART) regimen con-

tainingDRV/r 600/100mg bid is estimated to be a cost-effective therapy when

compared with a HAART regimen containing LPV/r, for the management of

treatment-experienced, PI-resistant, HIV-infected adults with a broad range

of previous PI use/failure.

Introduction

Since the mid-1990s, the widespread use ofhighly active antiretroviral therapy (HAART) inthe USA and other industrialized countries hasresulted in a significant extension of the life ex-pectancy of HIV-infected patients.[1-4]

Despite these major therapeutic advances,there is still no cure for HIV. In addition, a sub-stantial proportion of people living with HIV orAIDS eventually fail and/or discontinue their in-itial HAART regimens. These patients will oftendevelop antiretroviral resistance in the process,potentially jeopardizing further treatment optionsas a result of the cross-resistance between cur-rently available antiretroviral agents; in particular,resistance to non-nucleoside reverse transcriptaseinhibitors (NNRTIs) and HIV protease inhibitors(PIs), which are the present-day cornerstones ofinitial HAART regimens.[5-7]

For the management of individuals with HIVinfection, HAART regimens containing the PI co-formulation of lopinavir with low-dose (200mg/day)ritonavir (LPV/r) are currently considered the stan-dard of care[8-10] because LPV/r has shown a higherlevel of efficacy than other so-called ritonavir-

boosted, PI-based regimens in randomized trialsin patients with varying degrees of antiretroviralexperience.[11-15]

Recently, Madruga et al.[16] published the48-week results of the phase III TITAN trial(TMC114-C214; TMC114/r In Treatment-expe-rienced pAtients Naive to lopinavir). This on-going trial compared the efficacy of HAARTregimens containing the recently approved ritonavir-boosted PI darunavir (DRV/r) at a daily dose of600/100mg twice daily (bid) with that of LPV/r-based HAART in treatment-experienced, LPV/r-naive, HIV-infected adults. After 48 weeks oftherapy, the TITAN results showed that sig-nificantly more patients treated with DRV/r hadplasma HIV-RNA levels less than 50 copies/mL,the current recommended goal of antiretroviraltherapy (ART), than patients treated with LPV/r(71% for DRV/r vs 60% for LPV/r; p= 0.005).

The higher response with DRV/r was observedin many patient subgroups, in particular amongthose TITAN patients who exhibited one or moremutations in the HIV protease gene at baselinefrom the set of primary PI resistance-associatedmutations (RAMs) published in the Internation-al AIDS Society – USA (IAS–USA) 2006[17]

130 Brogan et al.

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

guidelines (72% for DRV/r vs 45% for LPV/r withplasma HIV-RNA levels less than 50 copies/mL;p= 0.003).[16,18,19] Mutations at these positions onthe HIV protease gene can lower the efficacy ofall approved PIs, especially if combinations ofthese mutations are present. Patients bearing atleast one of these mutations are therefore con-sidered to be exhibiting PI resistance and areoften challenging to treat because PIs are thecornerstone of HAART in treatment-experiencedpatients.[5,10,20]

The objective of this study was to determinewhether, from aUS societal perspective, a HAARTregimen including DRV/r 600/100mg bid is costeffective compared with a HAART regimen in-cluding LPV/r whenmeasured as incremental costper quality-adjusted life-year (QALY) gained intreatment-experienced, HIV-infected adults withevidence of PI resistance (i.e. having ‡1 IAS–USAprimary PI RAM).

Methods

Model Structure and Flow

A Markov cost-effectiveness model with a3-month cycle time, originally developed to assessthe cost effectiveness ofDRV/r 600/100mg bid usingthe results of the POWER 1 and 2 (TMC114-C213and C202; Performance Of TMC114/r When eval-uated in treatment-Experienced patients with pro-tease inhibitor Resistance) trials, was adapted toreflect the target population of the analysis and to useinput parameters specific to the one or more IAS–USAPImutation subpopulation of theTITAN trial.

The model[21] followed a cohort of treatment-experienced adults infected with HIV-1 who ex-hibited one or more IAS–USA primary PI RAMsat treatment initiation.[17] The model’s popula-tion was designed with a demographic profile,baseline CD4 cell count distribution and plasmaHIV-RNA level similar to those of the corre-sponding subpopulation in the pooled DRV/r600/100mg bid and LPV/r arms of the TITANtrial (table I).[16,19]

The model simulated disease progression overtime in this target population by modelling thepopulation’s evolution through different health

states characterized by CD4 cell count ranges(>500, 351–500, 201–350, 101–200, 51–100 and0–50 cells/mm3) and, eventually, the absorbingstate of death. These health states, or CD4 cellcount ranges, were chosen based on their clinicalrelevance in the management of HIV-infectedindividuals and the generally recognized pre-dictive power of an HIV-infected individual’sshort-term (e.g. 3–6-month) risk of clinical dis-ease progression.[9,10,17,22]

The model structure is presented in figure 1.Every 3 months participants either remained intheir current state or moved to any other state ofthe model, including death. The arrows in figure 1display the possible movements of participants withCD4 cell counts between 201 and 350 cells/mm3.Similar movements (not shown) are possible forparticipants in the other states of the model, exceptdeath, which is an absorbing state. Depending onthe model input parameters, the probability ofsome transitions may be zero.

A US societal perspective was used in themodel with direct medical costs in the numeratorand QALYs in the denominator. The indirect costsincluding productivity losses are assumed to becaptured in the QALY estimates as recommended

Table I. Population characteristics for all participants in the one or

more primary International AIDS Society – USA protease inhibitor

resistance-associated mutation subset from the TITAN trial[19]

Input parameter ‡1 IAS–USA PI mutation

subset from TITAN (%)

Gender

Male 86.3

Female 13.7

Age (years)

15–39 29.5

40–64 68.4

65+ 2.1

CD4 cell count range (cells/mm3)

0–50 8.0

51–100 7.4

101–200 20.7

201–350 29.3

351–500 19.7

>500 14.9

IAS–USA = International AIDS Society – USA; PI =protease inhibi-

tor; TITAN=TMC114/r In Treatment-experienced pAtients Naive to

lopinavir.

Cost Effectiveness of DRV/r vs LPV/r in the USA 131


for the reference case by Gold et al.[23] In eachCD4 cell count range, participants incurred ART-related costs as well as non-antiretroviral-relateddirect costs based on estimates of healthcare ser-vices used during that health state taken from apublished US costing study.[24] These healthcareservices included multidrug ART, opportunisticinfection, prophylactic and other outpatient drugs,hospital stays, emergency room visits and out-patient physician visits. Each CD4 cell countrange in the model also had an associated utilityvalue[25] that was used to convert the time spent ineach health state over an individual’s remaininglifetime into estimates of QALYs. The impacts ofgains in productivity because of reduced morbidityor premature mortality were assumed to be cap-tured in the QALYs-gained measure, as recom-mended for the ‘reference case’ cost-effectivenessanalysis.[23]

As there were no large-scale differences in ad-verse event (AE) profiles between the DRV/r and

LPV/r groups in the TITAN trial, the model didnot make specific adjustments to costs or qualityof life, to account for AEs beyond those accountedfor by the cost and utility estimates for the CD4cell count ranges.

Mortality rates were also assigned to eachCD4 cell count range in the model for both HIV-related and non-HIV-related deaths. HIV-relatedmortality rates were assumed to vary by CD4 cellcount range and were taken from published in-ternational cohort studies.[26] The non-HIV-relateddeaths were assumed to vary by age and weretaken fromUS life tables with a relative risk valueapplied to account for higher non-HIV-relatedmorality rates among individuals with HIV infec-tion than among those without such infection.[27]

The model assumed that, after starting a newtreatment, each individual will experience an in-itial period of rapidly increasing CD4 cell count,followed by a period during which the CD4 cellcount may vary more slowly or stabilize. Themodel also assumed a third period, characterizedby a decreasing CD4 cell count. This is in linewith the short- and long-term kinetics of CD4 cellcount usually observed after the initiation of ef-fective HAART, as well as with the progressiveimmune deficiency that eventually develops whenthese regimens no longer succeed in maintainingtheir inhibitory effect on viral replication.[28,29]

Themagnitude of the change in CD4 cell countover time, the duration of each of the above-mentioned periods of CD4 cell count kinetics,and the consequent transition probabilities be-tween health states in the model were assumed tobe dependent on the degree of the treatment-induced virological suppression and were pro-grammed separately in the model for differentcategories of viral response. The model definedthree levels of virological response to treatment,24 weeks after the start of a new regimen: viro-logical suppression below the limit of detection(HIV-RNA level <50 copies/mL), partial suppres-sion (between 50 copies/mL and a ‡1 log10 drop inplasma HIV-RNA level, referred to hereafter as50 copies/mL to ‡1 log10 drop), and no suppres-sion (<1 log10 drop in plasma HIV-RNA levelfrom baseline). These plasma HIV-RNA levelresponse categories were chosen because they

CD4201−350

CD4101−200

CD451−100

CD40−50

CD4 >500

CD4351−500

Death

Fig. 1. Schematic diagram of the cost-effectivenessmodel: illustrationof the transitions for individuals starting in the 201–350 cells/mm3

CD4 cell count range.

132 Brogan et al.


have been shown to correlate with both the mag-nitude of increase in the CD4 cell count and thelikely duration of response to treatment.[10,28,30,31]

This economic evaluation focused on the com-parison of the following initial treatment options:DRV/r 600/100mg bid and LPV/r 400/100mg bid.In line with the TITAN study design and currentstandard of care,[8-10,16] each regimen compared inthe economic evaluation also included an optimizedbackground regimen (OBR)made up of twoormorenucleoside reverse transcriptase inhibitors (NRTIs),with or without a NNRTI, chosen on the basis ofeach patient’s treatment history and resistance pro-file and the investigator’s opinion (see figure 2).

Eventually, individuals failed their initial reg-imen and switched to a follow-up regimen.Patients were assumed to remain on that regimenuntil death, despite declining CD4 cell counts. Inthe base case of this analysis, all individuals wereassumed to change to a HAART regimen con-sisting of TPV/r 500/200mg bid with an OBRconsisting of NRTIs with or without the fusioninhibitor enfuvirtide. This regimen was chosenbecause of the activity of TPV/r against PI-resistantstrains, including strains exhibiting reduced sus-ceptibility to DRV/r, and the long-term clinicalevidence gathered on TPV/r in highly treatment-experienced patients in the RESIST (RandomizedEvaluation of Strategic Intervention in multi-drug

reSistant patients with Tipranavir) trials.[32] In thebase-case analysis, the use of enfuvirtide in theswitch regimenwas assumed to be discontinued aftera 6-month period of decreasing CD4 cell count.

For each regimen considered by the model,clinical trial data[16,19,32-35] and data from otherpublished sources[28,31,36-40] were used to estimate:(1) the proportion of individuals in each virolog-ical response category after 24 weeks of therapy;(2) the duration and magnitude of the CD4 cellcount changes for each virological response cat-egory; and, by inference, (3) the 3-month transi-tion probabilities for movements between theCD4 cell count health states used in the model.

Model Input Parameters

A number of model inputs were derived fromthe original economic model.[21] These input pa-rameters (see table II), which are described indetail elsewhere, included the HIV-related andnon-HIV-related (not shown) mortality rates,utility values and direct non-antiretroviral-relatedcost of care estimates associated with each of themodel’s health states.[21,24-27] In addition, esti-mations of the patient’s management costs dur-ing the last 3 months of life, regardless of thecause of death, were taken from the originaleconomic model.

First regimen Switch regimen

Treatmentfailure

andswitcha

DRV/r

Darunavir 600 mg bid+

Ritonavir 100 mg bid+

OBR

LPV/r

Lopinavir 400 mg bid+


OBR

Death

TPV/r

Tipranavir 500 mg bid+


OBR

Fig. 2. Model structure, including primary comparisons. NOTE: OBR selection based on resistance testing; OBR may include NRTIs,NNRTIs, and a fusion inhibitor. Treatment failure defined as a decline in CD4 cell count. Switch may occur several time periods after the start ofCD4 cell count decline or before CD4 cell count decline. bid = twice daily; DRV/r = darunavir/ritonavir; LPV/r = lopinavir/ritonavir;OBR =optimized background regimen; NNRTI = non-nucleoside reverse transcriptase inhibitor; NRTI =nucleoside reverse transcriptaseinhibitor; TPV/r = tipranavir/ritonavir.



Efficacy-Related Inputs

Virological Response Rates at 24 Weeks

In addition to the above-mentioned demo-graphic profile, baseline CD4 cell count distribu-tion and baseline plasma HIV-RNA level, 24-weekvirological response rates for the initial LPV/rand DRV/r-based regimens were directly derivedfrom the clinical trial data gathered for therespective subgroups of TITAN patients withone or more primary IAS–USA PI mutations atbaseline.[16,19] For the TPV/r-based switch reg-imen, published data were used from the (pooled)RESIST 1 and 2 trials to estimate the plasmaHIV-RNA level outcomes at 24 weeks for TPV/r(see table III).[32-35]

Duration of Phases of CD4 Response

The duration of the different phases of CD4cell count kinetics during treatment with the in-itial and switch regimens considered by the modelare shown in table III. For the initial DRV/r andLPV/r-based regimens, the duration of the initialperiod of rapid CD4 cell count increases for eachof the 24-week virological response categories inthe model was based on the kinetics of the CD4cell count observed in the 48-week intention-to-treat analysis of the DRV/r and LPV/r arms in thesubgroup of TITAN participants with one ormore IAS–USA PI mutation at baseline. In theTITAN study, these patients experienced a veryrapid increase in CD4 cell count in the first 24weeks, followed by a slower CD4 cell count evo-lution between 24 and 48 weeks in both studyarms, and this slower evolution was irrespectiveof the degree of virological response at 24 weeks.Therefore, in the model, the duration of the

period of rapid, initial CD4 cell count increasefor DRV/r and LPV/r-based HAART was set to0.5 years in the base-case analysis for all virologi-cal response categories.

The base-case estimations of the duration ofthe period of more slowly changing CD4 cellcount or CD4 cell count stabilization (2 years) forpatients with plasma HIV-RNA levels of lessthan 50 copies/mL after 24 weeks of therapy onboth initial regimens were based on the results ofseveral long-term studies. These studies indicatedthat long-term treatment success and continuingincreases in the CD4 cell count over several yearswas possible for HIV-infected individuals inwhom HIV replication is rapidly suppressed be-low assay detectability, comparedwithHIV-infectedindividuals who did not show such suppression,irrespective of the degree of previous treatmentexposure.[30,36-40,43,44] From the same evidence, ashorter duration of the period of slowly evolvingCD4 cell count or CD4 cell count stabilization(0.5 years) was assumed for those patients whoexhibited an intermediate virological responseafter 24 weeks of therapy, because of the lowerprobability of achieving complete virologicalsuppression and the higher risk of prematuretreatment failure and discontinuation.[43,44] Incurrent medical practice, a less than 10-fold de-crease of the plasma HIV-RNA level belowbaseline levels after 24 weeks of treatment intreatment-experienced patients is generally re-garded as a treatment failure, warranting treat-ment modification.[9,10,17] This is particularly truefor patients with profiles similar to patients en-rolled in the TITAN trial, who generally had alesser degree of previous treatment exposure, less

Table II. HIV-related mortality rates, utility weights, and 3-month non-antiretroviral-related costs of care, in 2008 US dollars, associated with

each of the model health states

CD4 cell count range (cells/mm3) HIV-related mortality

(% per patient-year)[26]Utility weights[25] 3-month non-antiretroviral-

related costs:

other drugs (US$)[24,41]

3-month non-antiretroviral-

related costs: inpatient/outpatient (US$)[24,41]

0–50 17.6 0.781 2409 10 033

51–100 5.5 0.853 1582 4719

101–200 2.2 0.853 1582 4719

201–350 0.8 0.931 1010 3396

351–500 0.4 0.933 1010 3396

>500 0.4 0.946 822 2292

134 Brogan et al.


advanced disease characteristics and more avail-able treatment options than the highly treatment-experienced patient population enrolled in thePOWER trials.[16,45] Therefore, in the base-caseanalysis of this TITAN-based adaptation, patientson the initial DRV/r and LPV/r-basedHAART reg-imens with less than a 1 log10 decrease from base-line in the plasma HIV-RNA level after 24 weekswere switched to the follow-up regimen after24 weeks of treatment, as a result of the suboptimalefficacy of the initial regimen, and the duration ofthe last two periods of CD4 cell count kinetics wasset to zero. Based on the same rationale, patientswho exhibited a complete or intermediate virolog-ical response at 24 weeks were not assumed toremain on a virologically failing regimen beforeswitching to the follow-up regimen, and did notexperience the period of prolonged CD4 cell countdecline that characterizes the third period of CD4cell count kinetics included in the model.

The durations of the phases of initial, rapidCD4 cell count increase and slow CD4 cell count

variation or CD4 cell count stabilization for theTPV/r-based switch regimen, were set to equalthose of the two initial regimens for all virologicalresponse groups. Unlike patients on the initialregimens, however, patients were assumed to re-main on this switch regimen for their remaininglifetimes, because there was assumed to be noalternative regimen. It should be remembered,however, that although in this model, tipranavirwas considered the next therapy (as it was as-sumed to be the best option when consideringlopinavir and darunavir failures), subsequentdata derived from the TITAN trial have shownthat tipranavir should not necessarily be thenext option for darunavir failures because mostTITAN patients failing on darunavir therapy re-tained susceptibility to most available PIs.[46]

Calculation of Transition Probabilities: First andSwitch Regimens

The calculation of the transition probabilitiesbetween the different health states in the model

Table III. Virological and immunological efficacy estimates of the initial and switch regimen included in the model[19,28,31-40,42]

Virological Patients CD4 cell count dynamics

response at (%) Rapid increase Slow change/stable Decrease

24 weeks 3-month

changeaDuration

(years)

3-month

changeaDuration

(years)

3-month

changeaDuration

(years)

DRV/r 600/100mg bid +OBR

<50 copies/mL 77.3 48.23 0.5 12.45 2.0 N/A 0.0

50 copies/mL to ‡1log10 drop

11.4 65.77 0.5 20.95 0.5 N/A 0.0

<1 log10 drop 11.3 19.57 0.5 N/A 0.0 N/A 0.0

LPV/r 400/100mg bid +OBR

<50 copies/mL 50.5 57.62 0.5 9.10 2.0 N/A 0.0


21.5 62.42 0.5 -29.69 0.5 N/A 0.0

<1 log10 drop 28.0 0.27 0.5 N/A 0.0 N/A 0.0

TPV/r 500/200mg bid+OBRb

<50 copies/mL 23.9 38.25 0.5 0.00 2.0 -1.51 Remaining

lifetime


17.3 52.03 0.5 0.00 0.5 -1.51 Remaining

lifetime

<1 log10 drop 58.8 17.20 0.5 N/A 0.0 -1.51 Remaining

lifetime

a Mean 3-monthly change in CD4 cell counts (cells/mm3).

b Tipranavir +OBR was used as the switch regimen.

bid = twice daily; DRV/r =darunavir/ritonavir; LPV/r = lopinavir/ritonavir; N/A =not applicable; OBR =optimized background regimen;

TPV/r = tipranavir/ritonavir.



for each virological response group during theinitial period of rapidly increasing CD4 cell countwas based on the magnitude of the CD4 cellcount increase from baseline observed during thefirst 24 weeks of therapy in the TITAN trial (forthe initial DRV/r and LPV/r-based regimens) andpooled RESIST 1 and 2 trials (for the TPV/r-based switch regimen).[16,32] These 24-week CD4cell count increases were first linearly trans-formed to match the 3-month cycle length in-cluded in the model (see table III). Next, the meanvalues and the standard deviations of these 3-monthCD4 cell count increases were used to constructnormal distributions for these increases. Thesenormal distributions were applied to individualsassumed to be at the midpoint in each of thestarting CD4 cell count ranges in order to givetransition rate estimates from each starting rangeto all other CD4 cell count ranges.

To calculate the transition probabilitiesamong the different CD4 cell count ranges for thephase of stable or slowly variable CD4 cell countkinetics for the initial and switch regimens, indivi-duals were assumed to be uniformly distributedwithin each CD4 cell count range. In addition,all individuals were assumed to achieve a mean,3-month CD4 cell count gain, computed usinglinear interpolation from the rate of CD4 cellcount change observed between 24 and 48 weeksduring the TITAN and the pooled RESISTtrials.[16,19,32] For individuals with plasma HIV-RNA levels of less than 50 copies/mL after 24 weeksof therapy, this calculated 3-month CD4 cell countgain was applied throughout the 2-year period ofslow CD4 cell count increase. The impact of thisextrapolation beyond the 48-week trial data wastested by varying the magnitude of the CD4 cellcount gain using the 95% confidence intervals(CIs) from the trial estimates in sensitivity anal-ysis (figure 3) and by varying the duration oftime to treatment failure in variability analysis(table IV).

The transition probabilities for the period ofdeclining CD4 cell count for patients on theTPV/r-based switch regimen were calculated usingan equation derived from the results of theMulticenter AIDS Cohort Study.[47] The equa-tion estimated the relationship between the an-

nual rate of CD4 cell count decline and plas-ma HIV-RNA level in patients not taking ART.The annual rate of CD4 cell count decline =–21.3–33.5 log10 (HIV-RNA level‚1000), whichwas subsequently adjusted using recent data thatindicated a slower rate of CD4 cell count dec-line at each plasma HIV-RNA level when onHAART treatment).[21,31] This approach yieldedan estimated decline of 6.03 cells/mm3 per year,or 1.51 cells/mm3 per 3-month cycle while ontreatment. This rate of decline during the thirdphase of CD4 cell count kinetics on the TPV/r-based switch regimen was assumed to be thesame for all three virological response categories.To calculate the transition probabilities amongall the CD4 cell count ranges for the decliningstage, individuals were once again assumed to beuniformly distributed within each CD4 cell countrange, and all individuals in all the ranges againwere assumed to experience the mean CD4 cellcount loss each cycle period.

Costs Considered by the Model

Costs considered in the model included anti-retroviral drug costs and non-antiretroviral-relatedcosts, and are presented in 2008 US dollars.

The costs of antiretroviral regimens were cal-culated using a micro-costing approach, multi-plying the daily dosage of each antiretroviraldrug as used in the combination regimens withthe cost per unit of dosing. The mean daily costfor each of these drugs was estimated from therecommended doses in the guidelines publishedby the US Department of Health and HumanServices,[10] and unit costs were obtained fromMedispan’s Price-Chek PC[48] and AnalySource’s[49]

wholesale acquisition costs. The daily cost for dar-unavir 600mg bid was set at US$27.40. A weightedaverage daily cost per patient was then calculated foreach modelled combination regimen, based on theproportion of patients using each drug in that regi-men and the daily cost of that drug. For the initialDRV/r-based and LPV/r-based regimens comparedin the model, data on the proportional use of anti-retroviral drugs (DRV/r, LPV/r, NRTIs andNNRTIs) were derived from the subset of patientswith one or more IAS–USA primary PI mutations

136 Brogan et al.


at baseline in the TITAN trial.[19] For the TPV/r-based switch regimen, the weighted average dailycost was calculated from the listed price of the re-commended dose of TPV/r (500/200mg bid) andthe degree of use of enfuvirtide as observed in theRESIST 1 and 2 trials.[34] Because no information isavailable on NRTI use by patients receiving TPV/rin the RESIST 1 and 2 trials, the proportional use ofNRTIs was set to equal the average of the individualNRTI use observed in patients treated with dar-unavir 600/100mg de novo and the control PIs in the

pooled POWER 1 and 2 trials.[42] Weighted average3-month antiretroviral costs per patient for eachcombination regimen included in the model werethen computed (see table V). The estimated drugcosts did not include mark-up or dispensing fees.

As previously mentioned, the non-antiretroviral-related cost estimates associated with each of themodel CD4 cell count ranges (table II) were takenfrom the original model developed by Mauskopfet al.[21] Those cost estimates were derived froma US survey study[24] and inflated to 2008 US

$15 000 $20 000 $25 000 $30 000 $35 000

Rate of CD4 decline (−60.83, 0 cells/yr)

HIV-related mortality (95% CI)

Rate of slow CD4 increase (95% CI)

Lopinavir/r virological response (95% CI for <50 copies)

Relative risk of non-HIV-related mortality (95% CI)

Darunavir/r virological response (95% CI for <50 copies)

Non-ART medical costs (−/+ 20%)

Rate of rapid CD4 increase for switch regimen (95% CI)

Duration of rapid CD4 increase (−/+ 3 months)

Baseline CD4 distribution (95% CI for 0−50 group)

Rate of rapid CD4 increase for initial regimens (95% CI)

Utility values (95% CI)

Cost of initial regimens*

Switch virological response (95% CI for <50 copies)

Gender distribution (95% CI for males)

Age distribution (95% CI for 65+)

End-of-life costs (95% CI)

Use of enfuvirtide in switch regimen (95% CI)

Par

amet

er

Incremental cost per QALY gained

Sensitivity Run 1Sensitivity Run 2

Fig. 3. Tornado diagram graph for one-way sensitivity analysis.NOTE: *Run 1: cost based on all NNRTI use shifted to nevirapine; Run 2: costbased on all NNRTI use shifted to efavirenz. ART = antiretroviral therapy; CI = confidence interval; NNRTI =non-nucleoside reverse tran-scriptase inhibitor; QALY = quality-adjusted life-year.



dollars using the medical care component of theconsumer price index.[41] The costs included costsof non-antiretroviral medications and inpatientand outpatient costs for HIV disease monitoringand the management of complications of HIVinfection or its treatment. As individuals veryclose to death may not have been interviewed forthe study by Gebo et al.[24] a separate cost ofUS$27 098 in 2008 US dollars[50] was included toaccount for costs during the last 3 months of life,regardless of the cause of death.

Time Horizon

In the base-case analysis a lifetime horizon wasused. Results at a 5-year time horizon are alsoreported in variability analyses.

Discounting

The discount rate used for the costs and ben-efits was 3%, as recommended by Gold et al.[23]

The model was also run using undiscounted val-ues as part of a variability analysis.

Sensitivity and Variability Analyses

Extensive one-way and probabilistic sensitivityanalyses (PSA) were performed to address sam-pling uncertainty. Figure 3 displays the parame-ters (regimen efficacy, drug use, populationcharacteristics, utility, cost and mortality) as wellas the ranges/distributions used for these para-meters in the one-way sensitivity analyses. Forthe PSA, the parameters and ranges includedmirrored those used in the original model.[21]

These ranges and distributions were derived fromTITAN clinical trial data and published datasources. The results of the one-way sensitivityanalysis are presented in a tornado diagram(figure 3). For the PSA, 1000 Monte Carlo si-mulations were run and the results are presentedas a cost-effectiveness acceptability curve (figure 4).

A number of additional analyses were per-formed to assess the impact of variability in mod-elling assumptions and treatment patterns. Theseanalyses included variation of the time horizon ofthe model (2 years, 5 years and 10 years), discountrates (undiscounted, 5% and 6%), time until dis-continuation of enfuvirtide after CD4 cell countdecline begins in the switch regimen, and as-sumptions related to the duration of the initialtherapy (duration of the period of slow CD4 cellcount variation/CD4 cell count stabilization) in

Table IV. Results of variability analysis

Scenario Incremental cost per

QALY gained (US$)

Base-case analysisa 23 057

Time horizon (years)

2 14 191

5 6763

10 4919

Discount rate (%)

0 27 078

5 20717

6 19660

Time to enfuvirtide discontinuation after CD4 cell decline begins in

switch regimen (months)

3 23 115

12 22952

Total time on initial regimen before switchb

1 year for <50 copies/mL group,

0.5 year for ‡50 copies/mL groups

32858

7 years for <50 copies/mL group,

2.5 years for 50 copies/mL to

‡1 log10 drop group, 1 year for

<1 log10 drop group

21330

a The base-case values for the variables changed in the variability

analysis were as follows: time horizon of lifetime; 3% discount

rate; enfuvirtide discontinued 6 months after CD4 decline begins

in switch regimen; total time on initial regimen before switch:

2.5 years for less than 50 copies/mL group, 1 year for

50 copies/mL to 1 log10 or greater drop group, 0.5 year for less

than 1 log10 drop group.

b This analysis was conducted by varying the period of stable or

slowly variable CD4 cell count kinetics.

QALY = quality-adjusted life-year.

Table V. Three-month drug regimen costs in 2008 US

dollars[10,48,49]

Drug regimen Daily drug costs (US$) 3-Month drug costs (US$)

DRV/r+OBR 64.81 5918

LPV/r +OBR 56.33 5144

TPV/r+OBR 89.80 8200a

a Drug costs assume 27.1% of people received enfuvirtide (pooled

RESIST 1 and 2 data).

DRV/r =darunavir/ritonavir; LPV/r = lopinavir/ritonavir; OBR =opti-mized background regimen; RESIST =Randomized Evaluation of

Strategic Intervention in multi-drug resiStant patients with Tipranavir;

TPV/r = tipranavir/ritonavir.

138 Brogan et al.


the three virological response groups consideredby the model to reflect the impact of potentialdifferences in anticipated total treatment dura-tion resulting from either more or less aggressiveapproaches to therapy failure and modification(for those patient groups with an incomplete vi-rological response at 24 weeks) or a longer/shorterperiod of successful response to treatment (for thosepatients who achieved an undetectable plasmaHIV-RNA level at week 24).

Results

Base-Case Analysis

Table VI presents the results of the base-caseanalysis. Over a lifetime horizon, the base-caseanalysis predicted that the use of DRV/r-basedHAART would be associated with an increaseddiscounted life expectancy of 0.461 years and anincreased discounted quality-adjusted life ex-pectancy of 0.493 QALYs when compared withLPV/r-based combination therapy. ComparedwithLPV/r, treatment with a regimen containingDRV/r 600/100mg bid also is expected to beassociated with an increased proportion of timein the higher CD4 cell count ranges (>350 CD4cells/mm3) and decreased time in the lower CD4cell-count ranges (£350 CD4 cells/mm3).

The predicted total discounted lifetime costs werehigher for the DRV/r regimen than for the LPV/rregimen, partly because of higher antiretroviral drug

costs and partly because of the increased number ofyears alive (13.496 vs 13.035 years) during whichtreatment is needed. As a result, the base-case in-cremental cost-effectiveness ratio (ICER) of DRV/rcompared with LPV/r in TITAN patients withone or more IAS–USA primary PI RAMs wasUS$24 654 per life-year gained and US$23 057per QALY gained.

Sensitivity and Variability Analyses

The results of the one-way sensitivity analysesare reported in figure 3. The most influential pa-rameters in this uncertainty analysis were the rateof CD4 cell count change during the period ofdecline, the probability of HIV-related mortality,the rate of CD4 cell count change during theperiod of stable or slowly variable CD4 cell countkinetics, the percentage achieving HIV-RNAlevels of less than 50 copies/mL for those receivingDRV/r and LPV/r-based regimens, and the re-lative risk of non-HIV-related mortality over thegeneral population. The model results remainedbelow the commonly cited threshold of US$50000per QALY gained in all sensitivity analyses. Fig-ure 4 shows the results of probabilistic sensitivityanalysis for a lifetime horizon, indicating thatthere is a 0.754 probability that the cost-utilityratio will be below the threshold value.

The results of the variability analyses are pre-sented in table IV. Changes to the model timehorizon had the greatest impact on the modelresults, most notably at 10 years, which yieldedan ICER of US$4919. Reducing the total time oninitial therapy before switching produced thelargest cost-utility ratio among scenarios tested(US$32 986 per QALY gained compared withUS$23 057 per QALY gained in the base-caseanalysis). For all scenarios tested, the ICER re-mained below US$50 000 per QALY gained.

Discussion

Mauskopf et al.[21] and Moeremans et al.[51]

have evaluated the cost effectiveness of darunavir600/100mg bid using data derived from thephase IIb POWER trials in highly treatment-experienced patients with one or more IAS–USA

0

0.2

0.4

0.6

0.8

1.0

$10 000 $30 000 $50 000 $70 000 $90 000 $110 000

Incremental cost per QALY gained

Pro

babi

lity

Fig. 4. Cost-effectiveness acceptability curve for darunavir/ritonavirvs lopinavir/ritonavir. QALY =quality-adjusted life-year.



primary PI RAM, which supported the initialapproval of DRV/r 600/100mg bid by US andEuropean regulatory authorities.[16,45,52,53] Thefocus of the economic analysis reported in thispaper, however, is on the subset of HIV-infectedpatients included in the darunavir phase IIITITAN trial who had a similar degree of reducedPI susceptibility/resistance, but less advancedHIV disease and a broader range of previous PIexposure/failure (none to two or more PIs) atenrolment than those in the darunavir phase IIbPOWER trials.[16,18]

Our analysis indicates that, when used in patientsmeeting the profile of this TITAN subpopula-tion, DRV/r 600/100mg bid, in combination withother antiretroviral agents, is expected to provideimproved life expectancy and reduced rates of dis-ease progression compared with LPV/r-containingHAART regimens. The DRV/r 600/100mg bidregimen is thus expected to lead to improvedquality-adjusted life expectancy in this popula-tion. In addition, patients are predicted to spendless time in the lower CD4 cell count ranges withthe DRV/r regimen than with the LPV/r regimen.

This is important not only for individual well-being, but also for the many other costs associatedwith HIV infection that are not included in thisanalysis. For example, productivity losses are moresubstantial once a person starts to develop oppor-tunistic infections.[54,55] This generally happens onceCD4 cell counts fall below 200 cells/mm3.[56,57]

When the difference in quality-adjusted survi-val was integrated with the expected difference intotal costs (but excluding the above-mentionedindirect costs) between DRV/r and LPV/r,DRV/r 600/100mg bid is estimated to be a cost-effective therapy, with a quality-adjusted ICERof US$23 057 per QALY gained in the base-caseanalysis. This ratio is well below US$50 000 perQALY gained, the commonly cited cost-effec-tiveness threshold value for the USA.

The cost effectiveness of DRV/r 600/100mgbid is robust to changes in modelling assumptionsand input parameter values. The probabilisticsensitivity analysis demonstrates that, with athreshold value of US$50 000 per QALY gained,treatment with DRV/r-based HAART rather thanLPV/r-containing regimens is a more favourable

Table VI. Base-case results for lifetime cost-utility analysis of darunavir/ritonavir compared with lopinavir/ritonavir

Outcome measuresa DRV/r LPV/r Difference

Life-years 13.496 13.035 0.461

QALYs 12.512 12.020 0.493

Mean time in each CD4 cell count range (cells/mm3) over remaining lifetime (years)

0–50 0.182 0.321 -0.139

51–100 0.310 0.438 -0.127

101–200 1.118 1.364 -0.245

201–350 2.862 3.007 -0.145

351–500 3.552 3.313 0.239

>500 5.471 4.593 0.879

Lifetime costs (US$)

Antiretroviral therapy 350 842 335 388 15455

Other drugs 54 681 55 113 -432

Inpatient care 104 601 109 027 –4425

Outpatient care 66 937 65 805 1131

End-of-life care 15 962 16 333 -370

Total 593 024 581 666 11358

Cost per life-year gained (US$) 24 654

Cost per QALY gained (US$) 23 057

a Discounted at 3% per year.

DRV/r =darunavir/ritonavir; LPV/r = lopinavir/ritonavir; QALY =quality-adjusted life-year.

140 Brogan et al.


choice in more than 75% of the simulations forthe lifetime time horizon. An important conclu-sion from the variability analyses performed isthat the cost-effectiveness ratio is very sensitive tothe time horizon chosen for the analysis. Shortertime horizons, which are clinically relevant giventhe average duration of therapies in treatment-experienced patients, produced substantially lowercost-effectiveness ratios than the base-case life-time time horizon. Treatment with a DRV/r reg-imen was shown to be cost effective comparedwith LPV/r-based combination therapy in allvariability and one-way sensitivity analyses.

Direct comparisons between the results of thisanalysis and those of economic studies of other PIoptions for treatment-experienced patients aredifficult. Simpson and colleagues[58] have publisheda US cost-effectiveness estimate of US$56 315 perQALY gained for TPV/r when compared withcurrently available PIs (excluding DRV/r).[58]This estimate was based on the results of theRESIST 1 and 2 trials, however, which enrolled apatient population with more advanced diseasecharacteristics and treatment experience, morecomparable to the highly treatment-experiencedpatients enrolled in the POWER trials than tothe patients modelled in our analysis.[32,45,59]

Simpson et al.[60] also recently published a cost-effectiveness study of HAART regimens includingLPV/r compared with ritonavir-boosted ataza-navir in treatment-experienced patients. Thislatter study used data from the BMS-045 trial inwhich LPV/r-based therapy appeared to dom-inate ritonavir-boosted atazanavir HAART over5 and 10 years. Unfortunately, this analysis didnot specifically report results for PI-resistantpatients and its relevance to the conclusions ofour analysis is therefore limited.[60]

Our model is based on the efficacy data gath-ered in a direct, head-to-head comparison ofLPV/r and DRV/r, as well as relevant publishedsources. It simulates the progression of the targetpopulation through the initial and follow-up re-gimens and is based on the strong confirmed linkbetween the level of viral suppression (mirroredin the plasma HIV-RNA level) and the durationof successful therapy. Moreover, like many othercost-effectiveness analyses of antiretroviral drugs

published in the HAART era, it uses the estab-lished relationship between the evolution of theCD4 cell count over time and the actual short-term risk of clinical disease progression[22,31] topredict clinical outcomes (and related costs) forthe target population. Another major strength ofour model is that it is designed to be flexible andallows for extensive sensitivity and scenario analyses.These analyses in turn allow for interpretation ofthe importance of uncertainties related to efficacyand other trial-based parameters, as well as theimpact of varying model assumptions and prac-tice patterns on the cost-effectiveness results.

Our model, like all others, has a number oflimitations. For instance, our model does notaccount for drug-specific costs or health effectsresulting from drug-specific AEs other than thoseincluded in the costs associated with each of theCD4 cell count ranges considered by the model.In the TITAN study, DRV/r 600/100mg gen-erally had a favourable safety and tolerabilityprofile that was broadly comparable to that ofLPV/r. There was, however, a considerably lowerincidence of grade 2–4 diarrhoea, an importantand troublesome side-effect of currently availablePIs, at least possibly related to study medicationin the DRV/r arm (8%) compared with the LPV/rarm (15%). Rash-related AEs, on the other hand,were observed more frequently in DRV/r treatedpatients (16%) than in patients receiving LPV/r-based therapy (7%). All these events were ofmild or moderate severity, except for two severeevents, which resolved without sequelae. Further-more, discontinuations because of diarrhoea andliver and lipid abnormalities were more frequentin the LPV/r group.[16,61,62] Therefore, the addi-tion of specific AE-related cost and outcomecomponents to the model would not be likely toaffect our conclusions regarding the cost effec-tiveness of DRV/r in the target population con-sidered in our analysis.

In the model we also assume that patients onthe initial DRV/r and LPV/r-based therapies even-tually switch to the same TPV/r-based follow-upregimen and continue on this regimen for theirremaining lifetimes. Recent virology data fromthe TITAN trial have, however, shown that mostdarunavir failures in TITAN retained susceptibility



to other available PIs, so the use of tipranavir isnot necessarily the next option for these pa-tients.[46] We chose this approach to keep the focusof our analysis on the two PI options comparedin the TITAN trial, and because of uncertaintyabout the evolution in treatment practices intreatment-experienced patients resulting from therecent advent of a number of new antiretroviralmedications, such as etravirine (a second-generationNNRTI), raltegravir and maraviroc (the proto-types of the new classes of HIV integrase andentry inhibitors, respectively).[63-66] This assump-tion nevertheless may be regarded as a limitationin themodel structure for twomain reasons. First,thanks to the advent of the above-mentioned newantiretroviral medications, the number of treatmentoptions for treatment-experienced patients failingtherapy – even those who are highly treatmentexperienced – has increased significantly. Second,in the TITAN study, the development of new PIand NRTI RAMs and the loss of susceptibilityto PIs or NRTIs were observed more frequentlyamong patients failing therapy in the LPV/r groupthan in patients failing therapy in the DRV/rgroup.[16,46] Because of this more prominent lossof PI and NRTI susceptibility in the LPV/r armand because of the treatment antecedents of themodelled patient population, the chances of build-ing a new HAART regimen using only PIs andNRTIs and at least two active agents (three activeagents are recommended by a number of treat-ment guidelines to ensure long-term treatmentsuccess)[9,10] will probably be smaller in the caseof failure on a LPV/r-based regimen than in thecase of failure on aDRV/r-based regimen. Becauseof the model’s switch of patients on DRV/r andLPV/r to a TPV/r-based follow-up combinationtherapy with an identical composition, our analysisdoes not account for the short- and long-term im-plications regarding costs. In addition, our analysisdoes not take into account the potentially delayedneed for more complex and expensive follow-up re-gimens (such as those containing maraviroc, ralte-gravir, or the injectable fusion inhibitor enfuvirtide)in TITAN-like patients failing DRV/r, when com-pared with LPV/r.[63-67]

Despite the above limitations, the conclusionsof our analysis further confirm the findings from

the POWER-based cost-effectiveness studies re-ported elsewhere that DRV/r is a cost-effectivetherapy in patients with PI resistance (i.e. one ormore IAS–USA PI RAMS).[21,51] These findingsare important because overcoming PI resistanceis one of the main challenges in the manage-ment of treatment-experienced, HIV-infected in-dividuals.[5,9,10,20]

It should be noted, however, that the TITANsubpopulation modelled in this analysis had lessadvanced HIV disease compared with the popu-lation enrolled in the POWER trials, as shown bythe substantially lower percentage of patients(15.4% vs 38.4%) with advanced immune defi-ciency (<100 CD4 cells/mm3).[21] Furthermore,only 60% of the TITAN subpopulation modelledin this economic analysis had been treated withthe three main antiretroviral classes currentlyavailable (PIs, NNRTIs and NRTIs) before en-rolment in the TITAN study, in contrast to thehighly treatment-experienced population in thePOWER trials, who were all at least three-classexperienced.[19,45]

More importantly, whereas all the patients inthe POWER trials had received (and failed) atleast two PI-containing HAART regimens beforeenrolment, 42% of the patients in the TITANsubgroup modelled in this analysis had receivedno PI or one PI before enrolment. This findingis in line with the knowledge that PI resistancecan develop in patients on first-line, PI-basedHAART when HIV is allowed to replicate in thepresence of the suboptimal PI drug pressuregenerated by a failing PI-based regimen.[5] It alsohas implications for decision-makers because itpoints to the value of DRV/r 600/100mg bid inthe management of treatment-experienced patients,even in those with limited previous exposure to,or failure on, PI-based regimens.

In addition to the currently approved 600/100mgbid dose, the phase III development programmeof DRV/r is performing an evaluation of the effi-cacy, safety and tolerability of a lower, once-dailydose ofDRV/r 800/100mg in treatment-experiencedpatients without darunavirRAMs (TMC114-C229ClinicalTrials.gov ref NCT00524368). These pa-tients accounted for 82% of the overall popula-tion included in the TITAN trial, which included

142 Brogan et al.


treatment-experienced patients with a broad pro-file similar to that usually encountered in theclinical setting.[16,18] Furthermore, the character-istics of these patients are in line with the pre-valence data in the overall population showing thatmost treatment-experienced patients have no dar-unavir RAMs.[68-70] This lower DRV/r dose hasshown impressive efficacy results and a favour-able tolerability profile in antiretroviral-naivepatients[71] and in the subgroup of POWERpatients with no darunavir RAMs. In this sub-group of patients from the POWER trials, theonce-daily DRV/r 800/100mg dose generated avirological response that was comparable to thatobserved with 600/100mg bid (62% vs 67%; dif-ference 5; 95% CI -23 to +30) and both regimenswere superior to control PIs (11%; p < 0.0001).[19]Less frequent dosing, in particular once-dailydosing, and lower pill burden have been shown toimpact adherence to ART and thereby improvelong-term treatment outcomes when comparedwith antiretroviral drugs that have more complexdosing schedules.[72-74] Further studies are neededto evaluate the cost effectiveness of this lower,once-daily dose of DRV/r in the majority oftreatment-experienced, HIV-infected adults oncethe trial results become available.

Conclusions

In summary, this analysis of the PI-resistantsubpopulation in the TITAN trial demonstratesthat, when compared with a HAART regimencontaining LPV/r, a HAART regimen contain-ing DRV/r 600/100mg bid may be a cost-effectivetherapy from a US societal perspective for treat-ment-experienced, PI-resistant, HIV-infected adultswith a broad range of previous PI exposure/failure.

Acknowledgements

The authors are grateful to Tony Vangeneugden, Ben VanBaelen, Els De Paepe, Eric Lefebvre, Sabrina Spinosa-Guzman,Frank Tomaka, Piet DeDoncker and the rest of the darunavirstudy team at Tibotec Pharmaceuticals, Mechelen, Belgium,for their contributions in analysing and generating theTITAN-specific efficacy data included in this analysis. The

authors also acknowledge Catherine Elliott (medical writer,Gardiner-Caldwell Communications, Macclesfield, UK) forher editorial support. Finally, they thank the study inves-tigators, the patients and their families for their participationand support during the TITAN trial. This project was sup-ported financially by Johnson & Johnson PharmaceuticalServices.

AB, JM and SET, employees of RTI Health Solutions,have received grant support from Janssen-Cilag, the manu-facturer of darunavir, to assist with the preparation of themanuscript. They were not restricted by Janssen-Cilag in theiranalysis or in their interpretation of the final results. ES is anemployee of Johnson & Johnson Pharmaceutical Services,Beerse, Belgium, and owns stock options and shares in thiscompany.

References1. Braithwaite RS, Justice AC, Chang CC, et al. Estimating the

proportion of patients infected with HIV who will die ofcomorbid diseases. Am J Med 2005; 118: 890-8

2. Hutchinson AB, Farnham PG, Dean HD, et al. The eco-nomic burden of HIV in the United States in the era ofhighly active antiretroviral therapy: evidence of continuingracial and ethnic differences. J Acquir Immune Defic Syndr2006; 43: 451-7

3. 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

4. Walensky RP, Paltiel AD, Losina E, et al. The survivalbenefits of AIDS treatment in the United States. J InfectDis 2006; 194: 11-19

5. Clavel F, Hance AJ. HIV drug resistance. N Engl J Med2004; 350: 1023-35

6. Mocroft A, Ledergerber B, Viard JP, et al. Time to vi-rological failure of 3 classes of anti-retrovirals after initia-tion of highly active antiretroviral therapy: results from theEuroSIDA study group. J Infect Dis 2004; 190: 1947-56

7. Sabin CA, Scullard G, Easterbrook P, et al. Treatment ex-haustion of highly active antiretroviral therapy (HAART)among individuals infectedwithHIV in theUnitedKingdom:multicentre cohort study. BMJ 2005; 330: 695-9

8. European AIDS Clinical Society (EACS). 2009 guidelinesfor the clinical management and treatment of HIV infectedadults in Europe; 2009 [online]. Available from URL:http://www.europeanaidsclinicalsociety.org/guidelinespdf/EACS-EuroGuidelines2009FullVersion.pdf [Accessed 2009Dec 10]

9. Gazzard B, Bernard AJ, Boffito M, et al. British HIV As-sociation guidelines for treatment of HIV-infected adultswith antiretroviral therapy, 2006. HIV Med 2006; 7: 487-503

10. US Department of Health and Human Services, Panel onAntiretroviral Guidelines for Adults and Adolescents(USDHHS). Guidelines for the use of antiretroviral agentsin HIV-1-infected adults and adolescents; 2008 Jan 29[online]. Available from URL: http://AIDSinfo.nih.gov[Accessed 2009 Dec 10]

11. DeJesus E, LaMarca A, Sension M, et al. The CONTEXTstudy: efficacy and safety of GW433908/r in PI-experiencedsubjects with virological failure (24 week results) [abstract



178]. 10th Conference on Retroviruses and OpportunisticInfections; 2003 Feb 10–13; Boston, MA, USA

12. Dragsted UB, Gerstoft J, Youle M, et al. A randomized trialto evaluate lopinavir/ritonavir versus saquinavir/ritonavirin HIV-1-infected patients: the MaxCmin2 trial. AntivirTher 2005; 10: 735-43

13. Eron J Jr, Yeni P, Gathe J Jr, et al. The KLEAN study offosamprenavir–ritonavir versus lopinavir–ritonavir, eachin combination with abacavir–lamivudine, for initial treat-ment of HIV infection over 48 weeks: a randomised non-inferiority trial. Lancet 2006; 368: 476-82

14. Johnson MA, Grinsztejn B, Rodriguez C, et al. Atazanavirplus ritonavir or saquinavir, and lopinavir/ritonavir inpatients experiencing multiple virological failures. AIDS2005; 19: 685-94

15. Walmsley S, Bernstein B, KingM, et al. Lopinavir–ritonavirversus nelfinavir for the initial treatment of HIV infection.N Engl J Med 2002; 346: 2039-46

16. Madruga JV, Berger D, McMurchi M, et al. Efficacy andsafety of darunavir/ritonavir compared with that of lopi-navir/ritonavir at 48 weeks in treatment-experienced, HIVinfected patients in TITAN: a randomized controlledphase III trial. Lancet 2007; 370: 49-58

17. International AIDS Society USA (IAS–USA). Treatmentfor adult HIV infection: 2006 recommendations of the In-ternational AIDS Society – USA panel. JAMA 2006; 296:827-43

18. 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

19. Tibotec Pharmaceuticals. Results of TITAN Phase III clin-ical trial for darunavir/r. City: Mechelen, Belgium. TibotecPharmaceuticals. Data on file, 2007

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

21. Mauskopf J, Brogan A, Martin S, et al. Cost effectivenessof darunavir/ritonavir in highly treatment-experienced,HIV-1-infected adults in the USA. PharmacoEconomics2010; 28 Suppl. 1: 83-105

22. Phillips AN, Lundgren JD. The CD4 lymphocyte count andrisk of clinical progression: the T cell in HIV infection anddisease. Curr Opin HIV AIDS 2006; 1: 43-9

23. Gold MR, Siegel JE, Russell LB, et al. Cost-effectiveness inhealth and medicine. New York: Oxford UniversityPress, 1996

24. GeboK, Fleishman J, Conviser R, et al. Contemporary costsof HIV health care in the HAART era [abstract 537]. 13thConference on Retroviruses and Opportunistic Infections;2006 Feb 5–9; Denver, Colorado, USA

25. Simpson KN, LuoMP, Chumney E, et al. Cost-effectivenessof lopinavir/ritonavir versus nelfinavir as the first-linehighly active antiretroviral therapy regimen for HIV infec-tion. HIV Clin Trials 2004; 5: 294-304

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

27. Jensen-Fangel S, Pedersen L, Pedersen C, et al. Low mor-tality in HIV-infected patients starting highly active anti-

retroviral therapy: a comparison with the general popula-tion. AIDS 2004; 18: 89-97

28. Deeks SG, Barbour JD, Grant RM, et al. Duration andpredictors of CD4 T-cell gains in patients who continuecombination therapy despite detectable plasma viremia.AIDS 2002; 16: 201-7

29. Kaufmann GR, Furrer H, Ledergerber B, et al. Character-istics, determinants, and clinical relevance of CD4 T cellrecovery to <500 cells/microl in HIV type 1-infected in-dividuals receiving potent antiretroviral therapy. Clin In-fect Dis 2005; 41: 361-72

30. Benzie AA, Bansi LK, Sabin C, et al. Viral rebound inpatients on antiretroviral therapy with viral suppression:association with extent of previous virological failure andtime with viral suppression [abstract O9]. 12th AnnualConference of the British HIV Association; 2006 29 Mar–1Apr; Brighton, UK

31. Ledergerber B, Lundgren JD,Walker AS, et al. Predictors oftrend in CD4-positive T-cell count and mortality amongHIV-1-infected individuals with virological failure to allthree antiretroviral-drug classes. Lancet 2004; 364: 51-62

32. Hicks CB, 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 inMulti-drug Resistant Patients with Tipranavir (RESIST) studies:an analysis of combined data from two randomised open-label trials. Lancet 2006; 368: 466-75

33. Cahn P, and the RESIST 2 Study Team. 24-Week data fromRESIST 2: Phase 3 study of the efficacy and safety ofbackground therapy plus tipranavir/ritonavir (/r) or opti-mized ritonavir-boosted standard-of-care (SOC) com-parator PI (CPI) in a large randomized multicenter trial intreatment-experienced HIV+ patients [abstract PL14.3].7th International Congress on Drug Therapy in HIV In-fection; 2004 Nov 14–18; Glasgow, UK

34. Cooper D, Hicks C, Cahn P, et al. 24-Week study analyses:the efficacy of tipranavir/ritonavir (TPV/r) is superiorto lopinavir/ritonavir (LPV/r), and the TPV/r treatmentis enhanced by inclusion of genotypically active anti-retrovirals in the optimized background regimen (OBR)[abstract 561]. 12th Conference on Retroviruses and Op-portunistic Infections; 2005 Feb 22–25; Boston, Massa-chusetts, USA

35. Hicks CB and the RESIST-1 study team. A phase 3 ran-domized, controlled, open-label multicenter trial comparingtipranavir/ritonavir (TPV/r) to an optimized comparatorprotease inhibitor/r (CPI/r) regimen in antiretroviral(ARV) experienced patients: 24-week data [abstract 3726].44th Interscience Conference on Antimicrobial Agents andChemotherapy; 2004 Oct 31–Nov 2;Washington DC, USA

36. Garcia F, de Lazzari E, Plana M, et al. Long-term CD4T-cell response to highly active antiretroviral therapy ac-cording to baseline CD4 T-cell count. J Acquir ImmuneDefic Syndr 2004; 36: 702-13

37. Hunt PW, Deeks SG, Rodriguez B, et al. Continued CD4cell count increased in HIV-infected adults experiencing4 years of viral suppression on antiretroviral therapy.AIDS 2003; 17: 1907-15

144 Brogan et al.


38. Kaufmann GR, Perrin L, Pantaleo G, et al. CD4 T-lym-phocyte recovery in individuals with advanced HIV-1 in-fection receiving potent antiretroviral therapy for 4 years:the Swiss HIV Cohort Study. Arch Intern Med 2003; 163:2187-95

39. Smith CJ, Sabin CA, Lampe FC, et al. The potential forCD4 cell increases in HIV-positive individuals who controlviraemia with highly active antiretroviral therapy. AIDS2003; 17: 963-9

40. Tarwater PM, Margolick JB, Jin J, et al. Increase and pla-teau of CD4 T-cell counts in the 3(1/2) years after initiationof potent antiretroviral therapy. J Acquir Immune DeficSyndr 2001; 27: 168-75

41. US Department of Labor, Bureau of Labor Statistics. USMedical Care Consumer Price Index; 2008 [online]. Avail-able from URL: http://data.bls.gov/cgi-bin/surveymost?cu[Accessed 2009 Dec 10]

42. Tibotec Pharmaceuticals. Results of POWER 1 and 2 PhaseII clinical trials for darunavir/r. City: Mechelen, Belgium.Tibotec Pharmaceuticals. Data on file, 2005

43. Maggiolo F, Migliorino M, Pirali A, et al. Duration of viralsuppression in patients on stable therapy for HIV-1 infec-tion is predicted by plasma HIV RNA level after 1 monthof treatment. J Acquir Immune Defic Syndr 2000; 25: 36-43

44. Yamashita TE, Phair JP, Munoz A, et al. Immunologic andvirologic response to highly active antiretroviral therapyin the Multicenter AIDS Cohort Study. AIDS 2001; 15:735-46

45. 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

46. De Meyer S, Lathouwers E, Dierynck I, et al. Character-ization of virologic failure patients on darunavir/ritonavirin treatment-experienced patients. AIDS 2009; 23: 1829-40

47. Cook J, Dasbach E, Coplan P, et al. Modeling the long-termoutcomes and costs of HIV antiretroviral therapy usingHIV RNA levels: application to a clinical trial. AIDS ResHum Retroviruses 1999; 15: 499-508

48. Medispan. Price-check PC; [online]. Available from URL:http://www.medispan.com/index.aspx [Accessed 2009 Dec 10]

49. Analy$ource� Online; [online]. Available from URL: http://www.analysource.com [Accessed 2009 Dec 10]

50. Yang Z, Norton EC, Stearns SC. Longevity and health careexpenditures: the real reasons older people spend more.J Gerontol 2003; 58: S2-10

51. 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;28 Suppl. 1: 107-28

52. 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]

53. European Commission Community register of medicinalproducts for human use. PREZISTA; [online]. Availablefrom URL: http://ec.europa.eu/enterprise/sectors/pharma

ceuticals/documents/community-register/html/h380.htm [Ac-cessed 2009 Dec 10]

54. Liu GG, Yin DD, Lyu R, et al. Economic costs of HIVinfection: an employer’s perspective. Eur J Health Econ2002; 3: 226-34

55. Sendi P, Schellenberg F, Ungsedhapand C, et al. Pro-ductivity costs and determinants of productivity in HIV-infected patients. Clin Ther 2004; 26: 791-800

56. Ahsan N, Langhoff E. Immunopathogenesis of humanimmunodeficiency virus. Semin Nephrol 1998; 18: 422-35

57. Holkmann Olsen C, Mocroft A, Kirk O, et al. Interruptionof combination antiretroviral therapy and risk of clinicaldisease progression to AIDS or death. HIV Med 2007; 8:96-104

58. Simpson KN, Chumney ECG, Hicks CB, et al. Cost effec-tiveness of tipranavir in treatment-experienced HIVpatients in the USA [abstract PIN10]. 11th Annual Inter-national Meeting of the International Society for Pharm-acoeconomics and Outcomes Research; 2006 May 20–24;Philadelphia, Pennsylvania, USA

59. 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

60. Simpson KN, Jones WJ, Rajagopalan R, et al. Cost effec-tiveness of lopinavir/ritonavir compared with atazanavirplus ritonavir in antiretroviral-experienced patients in theUS. Clin Drug Invest 2007; 27: 443-52

61. Rockstroh J, Clumeck N, Spinosa-Guzman S, et al. Dar-unavir (TMC114)/r has tolerability and efficacy benefitsfor treatment-experienced patients compared with controlPIs; overview of the POWER trials [abstract P28]. 8th In-ternational Congress on Drug Therapy in HIV Infection;2006 Nov 12–16; Glasgow, UK

62. Sherman DS, Fish DN. Management of protease inhibitor-associated diarrhea. Clin Infect Dis 2000; 30: 908-14

63. Committee forMedicinal Products for HumanUse (CHMP)of the European Medicines Evaluation Agency (EMEA).Summary of positive opinion for raltegravir; 2007 Nov 27[online]. Available from URL: http://www.ema.europa.eu/pdfs/human/press/pr/52210907en.pdf [Accessed 2009Dec 10]

64. 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.ema.europa.eu/humandocs/Humans/EPAR/celsentri/celsentri.htm [Accessed 2009Dec 10]

65. 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 [Ac-cessed 2009 Dec 10]

66. 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 [Ac-cessed 2009 Dec 10]

67. Horne R, Clotet B, Cohen C, et al. Treatment-experiencedpatient perceptions of self-injectable therapy [abstract



PE7.3/25]. 10th European AIDS Clinical Society Con-ference; 2005 Nov 11–14; Dublin, Ireland

68. Loveday C,MacRae E. Susceptibility of a protease inhibitor(PI) treatment-experienced UK clinical cohort to TMC-114[abstract PL2.2]. 8th International Congress on Drug Ther-apy in HIV Infection; 2006 Nov 12–16; Glasgow, UK. 2006

69. Poveda E, de Mendoza C, Martin-Carbonero L, et al. Pre-valence of darunavir resistance mutations in HIV-1-infectedpatients failing other protease inhibitors. J AntimicrobChemother 2007; 60: 885-8

70. Mitsuya Y, Liu TF, Rhee SY, et al. Prevalence of darunavirresistance-associated mutations: patterns of occurrence andassociation with past treatment. J Infect Dis 2007; 196: 1177-9

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

72. 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.): S123-7

73. Maggiolo F, Ripamonti D, Gregis G, et al. Once-a-daytherapy for HIV infection: a controlled, randomized studyin antiretroviral-naive, HIV-1-infected patients. AntiviralTher 2003; 8: 339-46

74. 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 Josephine Mauskopf, Vice President,Health Economics, RTI Health Solutions, 3040 CornwallisRoad, Research Triangle Park, NC 27709, USA.E-mail: [email protected]

146 Brogan et al.


us cost effectiveness of darunavir/ritonavir 600/100mg bid in treatment-experienced, hiv-infected...

Documents