CENTER FOR DRUG EVALUATION AND
`
`RESEARCH
`
`APPLICA TION NUMBER:
`
`21-2 72
`
`MEDICAL REVIEW
`
`

`

`NDA 21~272 Remodulin ® (UT-15, Treprostinil) Protocol P01:l3 review 04/) 8/02 l:09 PM page l
`
`Division of Cardio-Renal Drug Products
`Medical Officer Review
`
`NBA 21-272 (serial # not submitted)
`RemodulinTM (treprostinil, UT-lS) injection
`Sponsor: United Therapeutics
`
`Date of Submission: 14 March 2002
`
`Reviewer: Abraham M. Karkowsky, M.D., Ph.D.
`
`'Date of Review: 14 March 2002
`
`Background: This is an amended protocol. UT-l 5 was approved under 21 CFR 314
`subpart H (314500-560). Approval was conditioned on the submission of a protocol that
`demonstrates an interpretable clinical benefit for UT—l 5. The Division met with United
`- Therapeutics and their consultants on 13 February 2002 during which the broad outline of
`a Flolan withdrawal study was explored. A protocol was received on 28 February 2002.
`Two major objections were raised and clarification of other issues was requested. The
`sponsor submitted a protocol on 13 march 2002. A meeting was held with the sponsor on
`7 March 2002. This reviewer was invited but unable to attend that meeting. The
`reviewer’s comments, however, were transmitted to those from the Agency who attended
`the meeting. Only cosmetic changes were recommended by the Agency. This reviewer,
`-
`however, still has substantial reservations related to the protocol. These reservations are
`listed at the end of the study.
`
`Protocol Review:
`
`Study number P01 :13.
`
`Title of Study: A Multicenter, Randomized, Parallel, Placebo-Controlled Study of the
`Safety and Efficacy of Subcutaneous RemodulinTM Therapy After Transition From
`Flolan® in Patients With Pulmonary Arterial Hypertension:
`
`Investigators and Sites: Not Specified.
`
`Formulations: 2.5, 5.0 and 10 mgml for continuous subcutaneous infusion or placebo.
`
`Inclusion criteria: Subjects that are enrolled are:
`0 Between 18-75 years.
`0
`If female incapable of childbearing.
`0 Have a diagnosis of pulmonary hypertension (either primary or secondary to systemic
`sclerosis syndrome).
`0 Class II or III status.
`
`I Whose status was stable for at least 30 days.
`0 Have a baseline walk distance of > 250 meters.
`
`

`

`NDA 21-272 Remodulin ® (UT-15, Treprostinil) Protocol P0] :13 review 04/18/02 1:09 PM page 2
`
`Receiving flolan at a dose of at least 20 ng/kg/min but less than 75 ng/kg/min.
`Received Flolan for at least 6 months and have been maintained on stable doses for at
`
`least 30 days.
`
`Unless contraindicated, be able to receive anticoagulants e.g. warfarin to achieve an
`INR of between 2.0 and 3.0 or heparin to produce an aP'IT of between 1.3 to 1.5 x
`control.
`
`Able to manage a subcutaneous pump.
`Stability of corticosteroid doses.
`
`Exclusion criteria: Subjects were excluded if:
`
`They are pregnant or nursing
`Had a new chronic therapy added for pulmonary hypertension or a stable medication
`changed within 30 days with the exception of anticoagulants.
`Received Remodulin or Bosentan (or other endothelial blocker) within 30 days.
`
`Have evidence of parenchymal lung disease. As indicated by:
`a) Total lung capacity < 60% predicted.
`b)
`W
`
`c) FEV/FVC ratio < 50%.
`d) If DLCO < 50% of that predicted, a high resolution CT must be performed to
`document diffuse interstitial fibrosis or alveolitis.
`
`HIV positive.
`Portal hypertension.
`Uncontrolled sleep apnea.
`Have a history of left sided heart disease including:
`a) aortic or mitral valve disease
`b) pericardial constriction or
`c)
`restrictive or congenital cardiomyopathy.
`Have evidence of current left-sided disease defined by:
`a) PCWP or left ventricular lefi-sided heart disease as defined by:
`b) LVEF < 40% by M(UGA or angiography or ECHO
`c) LV shortening of < 22% by ECHO
`d) Symptomatic coronary disease.
`Other disease (e.g. sickle cell disease) associated with pulmonary hypertension.
`Musculoskeletal disorder limiting ambulation.
`Uncontrolled hypertension (SBP > 160 or DBP > 100 mm Hg).
`Use of appetite suppressant within 3 months.
`Have chronic renal disease (Cr > 3.5 mg/dL).
`Recent investigational new drug or device.
`Have an atrial septostomy.
`Serious life-threatening disease.
`Unstable psychiatric status.
`Have anemia Hgb< 10 gm/dL
`
`Primary end point:
`
`
`
`

`

`NDA 21-272 Remodulin ® (UT-l5, Treprostinil) Protocol P01:l3 review 04/18/02 1:09 PM page 3
`
`The primary endpoint of the study is the time to clinical deterioration defined as
`the time from initiation of study drug to earliest incidence of clinical worsening of PAH
`symptoms requiring reinstitution of Flolan therapy or re-hospitalization or death.
`Any decision to re-institute Flolan should be supported by documented by objective
`criteria that the subject’s status has deteriorated despite attempts to increase the dose of
`the study drug or placebo. The preferred assessment criteria consist of the following
`parameters: PAH clinical status, 6~minute walk distance, Borg dyspnea score, dyspnea
`evaluation scale, transcutaneous 02 saturation, clinical signs and symptoms of PAH. If
`practical, the patient should be asked to perform light activity such as walking, to help in
`assessing whether clinical deterioration has occurred during the dose transition period.
`
`The study investigator is responsible for determining whether the subject’s status has
`deteriorated.
`
`An independent adjudication process will be utilized to assess all deterioration events as
`well as all withdrawals. Those patients, not weaned from Flolan at the end of the 2-week
`period, would be considered a treatment failure. Patients who withdraw due to reason
`other than clinical deterioration will be censored. The time to clinical deterioration will
`be compared between treatment groups using a proportional hazard regression model,
`adjusting for Flolan dose.
`
`Secondm end-points:
`0 Exercise capacity and Borg dyspnea score (assessed individually as well as through:
`and index composed of both these components).
`0 Dyspnea fatigue index.
`0
`Signs and Symptoms of PAH.
`0 Hospitalization for cardiovascular events or conditions.
`
`The walks at each week of testing (at the end of the transition period and weeks 4 and 8)
`will be fitted as a function of the initial Flolan dose and distance walked at baseline.
`Standardized mid—ranks will be calculated. Subjects who experienced clinical
`deterioration will be assigned a standardized rank of zero, with the rank carried forward
`to the week 8 value. Standardized ranks of the resulting values will be calculated. Similar
`analysis will be performed for the Borg dyspnea scale. An arithmetic average of the mid-
`ranks will be calculated for the combination of the 6-minute walk and Borg dyspnea.
`
`Changes in both measures will be assessed at Week 8 using a non-parametric analysis of
`covariance within the framework of the extended Cochran-Mantel Haenszel test.
`
`Statistical analyses:
`
`Two interim analyses are planned, limited to safety. An independent contractor will
`analyze the data for adverse events and deaths, with the analysis submitted to the DSMB.
`
`A sample size of approximately 90 patients would provide 90% powered using a two-
`sided log rank test at a level of 0.05 to detect a 2.8 fold increase in the median tome to
`
`
`
`

`

`NDA 21.272 Remodulin ® (UT—l S, Treprostinil) Protocol P01:l3 review 04/ 18/02 1:09 PM page 4
`
`clinical deterioration from a placebo median time of 4.6 weeks. In order to take in to
`account dropouts, the study size is increased to 100 subjects. The study will be continue
`until at least 50 clinical deterioration events have occurred.
`
`Randomization: Subjects will be randomized in a 1: 1 ratio and stratified by the current
`dose of Flolan (> 35 ng/kg/min and 5 35 ng/kg/min). The block sizes will be variable.
`
`Dosing: The dosing schedule for dose reduction of Flolan and the institution of UT-
`lS/placebo is shown below. The transition period is defined by days and not by specific
`hours.
`
`Table 1- Planned dose modifications for stu . POlzl3
`
`
`
`Im—-.—I Da #
`I--_I
`E-—_I
`-_-I
`l-_—I
`70% ofinitiaIFloIan Dose I
`_——I
`
`
`95% of initial Flolan Dose I
`
`
`
`Additional 5- l 0% as needed
`
`
`
`General guidelines to doses;
`0 The above dose should be followed as closely as possible.
`0 ‘ Increases in symptoms of PAH should be treated with increases in study drug first,
`even if it deviates from the above dosing recommendations.
`Should there be side effects suggesting an excessive effect, the dose of Flolan should
`be preferentially lowered.
`o A study dose/placebo increase should occur at least one hour before a corresponding
`decrease in the Flolan dose.
`'
`
`0
`
`o The subjects should not be discharged from the hospital until stable for at least 24
`hours after the dose of Flolan is stopped.
`If the subject could not be withdrawn from Flolan by the end of day 14, the subject
`would be considered a treatment failure.
`
`o
`
`0 No cardiac catheterizations should be conducted.
`
`Each subject is to be informed that infusion site pain is an expected outcome.
`
`The dose could be down-titrated for the following reasons
`0 Any measured or observed changes in vital signs or clinical signs that suggest an
`excessive drug effect.
`0 Any adverse experience possibly related to Remodulin e.g. headache, nausea,
`restlessness and anxiety.
`0 Onset of significant pain at the infusion site.
`
`Concomitant medications that were used prior to treatment are allowed.
`The listing of procedures during the study is shown below:
`
`
`
`

`

`NDA 2l-272 Rernodulin ® (UT~15, Treprostinil) Protocol P01 :13 review 04/18/02 1:09 PM page 5
`
`Table 2- list of urocedures durin; stud P01 :13
`
`
`
`
`
`
`
`.1
`
`K’
`Baseline
`_ reening‘
`U
`— ay—vtoo mo
`Informed consent' Inclusion/Exclusion criteria
`X
`
`Medical History fPE/Vital Signs/ 1 2-lead
`ECG/Labs
`
`
`
`
`
`
`ma
`
`
`
`
`
`
`
`
`b. May be less than 1-week if all procedures completed
`a May be performed up to one week prior to randomization
`c. Patient will return i 7 day seven even if prematurely discontinued. d. As needed for patient stability and prior to discharge
`
`e. A practice walk test should may be performed up to 6 weeks before randomization
`
`f. During the transition from Flolan to UT-l Sl'placebo the walk/Borg test may be performed periodically and as soon as possible for pre-
`mature termination
`
`
`h. Continuous monitoring during dosing and transition period
`g. Afier all baseline eligibility is determined
`i. Drug may be adjusted as outpatient transition period
`j. See text and table 1
`It Data should be collected immediately prior to
`ear discontinuation.
`
`
`
`
`Illlliiilll ”T”arellw'I
`
`1 da 1-14
`
`4 da 2s
`
`a da 56°
`
`.
`
`~-----—--—)
`
`i
`
`i
`
`
`
`
`
`
`
`Termination of study: The study can be terminated for the following reasons:
`0 The principal investigator or IRB elects to discontinue the study.
`FDA regulations are not observed.
`The protocol is violated.
`The data are of poor quality.
`Changes in personnel or facilities adversely effect performance of the study.
`
`Comments: This protocol is improved but not optimal. At the end there are likely several
`interpretations aside from a benefit for UT-l 5 for the treatment of puhnonary
`hypertension.
`
`0 There is little if any data to define how Flolan should be down-titrated except in the
`context of excessive hemodynamic effect. The current labeling for Flolan indicates
`that during clinical exposure there was a slow up~titration of Flolan at a rate of 2-3
`ng/kg/min every 3 weeks during stable treatment. It would seem that the rate of down
`titration should be the reverse mirror image of the up-titration scheme. The proposed
`protocol, however, proposes a much more rapid decrease in Flolan doses. For
`example, a subject on 35 ng/kg/min of Flolan at baseline would down titrated 5.2
`ng/kg/min on the first day, an additional 5.2 ng/kg/min on the second day and
`additional amounts on subsequent days and not 2-3 ng/kymin over 2-3 weeks. The
`down titration phase for Flolan might precipitate a withdrawal response.
`Consequently, any inference from this study if outcome events occurred during the
`flolan withdrawal could be interpreted as UT-l 5 mitigation of the effects of Flolan
`withdrawal as opposed to a'beneficial effect of UT—l 5 on puhnonary hypertension.
`
`With respect to the switch-over fi'om Flolan to UT-l 5, the sponsor in the original
`NDA submitted a preliminary description of Study P01:l l in which three subjects
`
`

`

`NDA 2l-272 Remodulin ® (UT-15, Treprostinil)Protocol P0] :13 review 04/l 8/02 l:09 PM page 6
`
`were transitioned from Flolan to UT-lS. Also submitted by the sponsor was an
`abstract1 in which 8 subjects were transitioned from, Flolan to UT-l 5 (I believe that
`the experience with the three subjects of study P01:11 are included in this abstract).
`The details are insufficiently described and the number of subjects too few to accept
`as known how to transition subjects off of Flolan to UTJ 5.
`
`Although there are provisions for additional decreases in the dose of Flolan if the
`hemodynamic effects are excessive because of the concurrent addition of UT—l 5,
`there are no provisions for decreasing the rate of down-titration if Flolan withdrawal
`appears to destabilize a subject. Nor is there the possibility to temporarily increasing
`Flolan until subject’s status stabilizes because of the rapid down-titration.
`
`Those subjects who destabilize are likely in the placebo control group if the UT-15
`acts to mitigate the withdrawal effects of Flolan. For example both Flolan and UT-15
`are peripheral vasodilators (not just pulmonary vascular vasodilators). Ifthc
`withdrawal of Flolan results in rebound peripheral hypertension, a second drug which
`vasodilates would apparently be interpreted as a useful drug for pulmonary
`hypertension but only serves to mitigate the peripheral effects of Flolan withdrawal.
`Given an inadequate database for defining the dynamic time course of Flolan
`withdrawal any deterioration during this portion of the study may not be easily
`interpreted as demonstrating a benefit for UT-I 5 during short-term withdrawal.
`
`Should a subject decompensate during the process of the switch over from Flolan to
`UT-lScontrol, the interpretation would therefore, be ambiguous. One potential
`inference is that UT-l 5 is an equivalent Flolan and is therefore, an active drug. The
`second interpretation is that UT-l 5 mitigated a rebound effect of the Flolan
`discontinuation and in itself is not active in the treatment of pulmonary hypertension.
`
`This reviewer would suggest that a much longer time frame be used for down-
`titration. In response to this concern the sponsor modified the crossover time to up to
`10 days from as short as < 2 days. Furthermore, this reviewer would allow up-
`titration of Flolan if acutely, the subject did not appear to tolerate the rapid down-
`titration. The sponsor, in the revised protocol, however, leaves no room for up-
`‘ titration of Flolan. In fact any increase in Flolan dose would be counted as a
`deterioration end point. This reviewer would use the firll two-week period for optimal
`down titration of Flolan as well as optimization of UT-l 5.
`
`The interpretation of this study as demonstrating efficacy would be markedly
`dependent on the outcome during this down—titration portion of the study. If nearly all
`events were a consequence of the Flolan down-titration the study could be interpreted
`as mitigating the withdrawal effects of Flolan. If the events occurred later in the study
`the interpretation would be less ambiguous.
`
`' Vachiery JL, Hill N, Zwicke D, Bars! R, Blackburn s, Naeije R International Society for Heart and Lung Transplantation 22"
`Annual Meeting and Scientific Session
`
`

`

`NDA 21-272 Remodulin ® (UT-15, Treprostinil) Protocol P01 :13 review 04/ 18/02 1:09 PM page 7
`
`0 Should the predominance of events occur during the Flolan withdrawal phase, an
`alternate interpretation of the outcome is that the UT-lS modifies hemodynamics and
`is useful for short term but not necessarily longer-term benefit. Since the purpose of
`UT-15 use is for chronic treatment, the importance of demonstrating a short-term
`benefit is unclear. For CHF, predominantly of left—sided origin, drugs are often
`classified for short term or chronic use. The anticipation is that drugs beneficial
`during the short term may not reflect long-term benefit to the patient. If all events
`occur during the Flolan withdrawal an argument can be made that all benefit is of
`hemodynamic/short term type.
`
`o The study results can be compromised by the consequence of site pain and the
`attendant unblinding.
`
`During the initial clinical studies (P01 :04 and P01 :05), subjects were started on a dose
`of UT—lS at an infiision rate of 1.25 ng/kg/min for the initial week then slowly up-
`titrated at subsequent weeks. Even with this conservative protocol, infiision site pain
`was nearly universal among those treated with UT-15. Pain or infusion site reaction
`listed as “severe” in intensity were each 39% (there could be overlap) and according
`to the sponsor lead to discontinuation in approximately 9% of those treated with UT—
`' 15. Using fairly aggressive dosing regimens would likely increase the rate of
`discontinuations in subjects. Should the dropout rate be too high due to infusion site
`pain the study may not be interpretable. In the two cases during the initial NDA
`where subjects were inadvertently started on fairly high doses of UT-l 5, both were
`hospitalized for intolerable pain. The dose of one of these two subjects was
`approximately 8.3 ng/kg/min. The dose at which the other subject was switched over
`is unclear. These doses would be achieved by approximately day 2 (presuming an
`initial Flolan of 35 ng/kg/min).
`
`Because of the occurrence of severe infusion site pain, there is the potential for
`informative censoring of subjects who discontinue. The protocol plans to censor any
`subjects who discontinue for reason other than worsening disease. Those who
`discontinue would likely not be allowed to remain untreated but would be re-started
`on Flolan. The question is how can one differentiate those that decided to discontinue
`because of intolerance to pain when compared to worsening or a non-improvement in
`pulmonary hypertensive status?
`
`In order to separate out the discontinuations for adverse events from those due to
`worsening status, as much data as possible would be collected on subjects at the time
`of discontinuation. The results would then be submitted to a blinded adjudication
`committee. Although this proposal is a marked improvement over the specifics in the
`initial protocol, the data submitted to the adjudication committee is filtered through
`the treating physician who may be entirely unblinded. The adjudicated resluts would
`therefore, not be entirely objective.
`
`3
`
`5
`
`é
`
`
`
`3
`
`
`
`3
`
`
`
`
`
`
`
`
`
`

`

`NDA 21-272 Remodulin ® (UT-15, Treprostinil) Protocol P01113 review 04/18/02 1:09 PM page 8
`
`This reviewer had suggested that a true intent-to treat study be performed, that Flolan
`not be restarted at the time of the discontinuation but only be re-started once
`symptoms of pulmonary hypertension supervene. The event would then be considered
`as an end-point. It appears to this reviewer that if it is ethical to discontinue subjects
`to placebo until symptoms reoccur it is equally ethical to discontinue the treatment till
`symptoms re-occur.
`
`It appears counter-intuitive that the exact same outcome i.e. terminating trial and
`restarting Flolan in one situation is treated as an outcome and in the other as a non-
`event.
`
`0
`
`It would be up to the sponsor to limit the number of subjects who discontinue. If none
`or few subjects discontinued, the study would be taken at its face value. If, however,
`many subjects discontinue, predominantly in the UT-l 5 group, the adequacy of the
`results based on censoring and imputation would not be convincing. The planned
`dosing regimen, in this context appears to be overly aggressive with high doses of
`UT-lS achieved over short period of times. Infirsion site pain with discontinuation is
`likely to be equal to or greater than observed in P01 :04 and P01 :05.
`
`o The primary endpoint of the study is the time to clinical deterioration defined as the
`time from initiation of study drug to earliest incidence of clinical worsening of PAH
`symptoms requiring reinstitution of Flolan therapy or re-hospitalization or death.
`Whereas death, assuming adequate follow-up, is objective, the other end points are
`subjective. The outcomes could be markedly altered by the degree of unblinding. It is
`likely that given the perceived benefit of UT-l 5, classification of events would be
`altered. The end-point of re-institution of Flolan would be aggresively pursued only
`in those without pain (placebo).
`
`The sponsor makes some effort to mitigate some of the consequence of unblinding.
`An adjudication process is to be performed by a committee blinded, to both treatment
`and the presence of infusion site pain. This analysis diminishes to a certain extent the
`degree of subjectivity in the outcomes. The problem, however, is the data, which is
`supplied to the adjudication committee is largely subjective and supplied by an
`unblinded (because of pain) observer.
`
`With respect to cause specific hospitalization as an end-point, any hospitalization, ‘
`aside from elective procedures or acute trauma, should be considered as an outcome.
`If those enrolled are relatively stable for 30 days prior to enrollment any
`hospitalization should be considered as worsening of disease.
`
`0 The adjudication committee’s charge is not defined. One way of defining a true
`discontinuation for adverse event versus deterioration, given the potential of
`unblinding and the subject nature of the endpoints, is to affirmatively require proof
`that those who discontinued for adverse events were not symptomatically worse. That
`is, absent adequate and convincing data the presupposition by the adjudication
`committee should be conservative and assume the subject deteriorated.
`
`
`
`

`

`NDA 21-272 Rernodulin ® (UT‘15, Treprostinil) Protocol P0]:13 review 04/18/02 l:09 PM page 9
`
`0 The overall safety of the study should be predicated on overall event rates that are
`lethal or potentially lethal. The sponsor purposefully does not include a Flolan group.
`The inherent assumption is that those stable on Flolan would unlikely deteriorate
`during this modest duration study. The charge to the DSMB is that the study should
`be discontinued based not on placebo versus UT-lS outcome but based on the
`outcome of the imputed Flolan group.
`
`0 The need for re-instituting Flolan once off treatment seems a reasonable surrogate for
`worsening of status. But the definition of “re-institution” needs some further
`discussion. This reviewer does not accept short—term increase in Flolan during the
`down-titration phase as indicating deterioration.
`
`0 The amended protocol limits inclusion to those on Flolan whose underlying disease is
`either primary pulmonary hypertension or for scleroderma spectrum of disease, which
`corresponds to the approved labeling of Flolan. Subjects whose pulmonary
`hypertension is due to congenital left-right shunting are now excluded. This reviewer
`concurs.
`
`0 Blinding may be compromised at the point when a subject crosses over at the end of
`8-weeks. There is no provision for re-titration or reinstitution of the original Flolan
`dose. Therefore, only placebo (vehicle) subjects will be re-titrated. The integrity of
`the data needs to be protected by this stage.
`
`0 Subjects with known left sided failure are to be excluded. There is no affirmative
`need to document the absence of lefi—sided cardiac disease. This reviewer would
`
`require affirmative proof that no left-sided disease exists prior to entry.
`
`0 An informed consent should prominently display the risk of stopping Flolan. In
`particular, there is some risk that should symptoms re~occur restarting Flolan may not
`salvage the subject if their status worsens and this might lead to death or
`hospitalization.
`
`0 The study limits the age range to < 18. During clinical studies children as young as 9
`years old were enrolled. Is there any reason for limiting the age?
`
`0 The sponsor should collect data on the need for and adequacy of pain medication.
`
`0 No CRFs were submitted.
`
`0 There are several walk tests to be performed during the flolan withdrawal phase of the
`study. It is unclear how these data would be used. Please clarify.
`
`0 The sponsor should clarify if those who discontinue due to site pain will be replaced.
`
`gonclusions:
`
`
`
`

`

`NDA 2l-272 Remodulin ® (UT-15, Treprostinil) Protocol P0] :13 review 04/l 8/02 1:09 PM page 10
`
`0 The study should be improved by allowing for intermittent up—titration of Flolan.
`Should two increases of drug/placebo doses not stabilize the subject a reversion of the
`last down—titration of Flolan should be allowed.
`
`0 The protocol should allow the full 2-weeks to optimize the down titration of Flolan.
`Not 8 days for down titration and six days for the optimization of either dose regimen.
`
`0 The definition for hospitalization should include any non-elective hospitalization
`
`0 The DSMB should be informed that the assumption is that there would be no or few
`events should the subject remain on Flolan. Safety should therefore be predicated on
`overall deaths, serious or irreversible events
`
`0
`
`-
`
`There is no stipulation that the DSMB can stop the study. Presumably this is implied
`in the mandate of the DSMB.
`
`If too many subjects discontinue due to site pain, the study is likely to be un-
`interpretable even with imputed values. The current dosing instructions for UT-l 5
`may be aggressive and may predispose to dropouts due to infusion pain.
`
`0 There are components of the end—points that are subjective. The information
`transmitted to the adjudication committee is likely to be colored by some knowledge
`as to treatment. Therefore, a small effect may not be convincing for efficacy.
`
`0- An effect driven by early dropouts in the placebo group could be interpreted as a
`mitigating effect of UT—l 5 on withdrawal.
`
`0 Alternatively, an effect driven by early dropouts could be interpreted as a benefit
`limited to short term therapy.
`
`0 The informed consent should prominently display the potential risk of death and the
`inability to easily salvage subjects upon cessation of Flolan. The current
`
`0 Blinding at the end of the 8-week period will be compromised since only those who
`were treated with placebo will require re-titration. The data needs to be locked in
`prior to unblinding or all patients should be restarted on their initial dose of Flolan.
`
`0 The CRFs should be submitted.
`
`0 The sponsor should capture need for pain medication. In addition to clarifying the
`down-side of UT—l 5 treatment, this parameter would be important in defining the
`adequacy of blinding.
`
`0 The adjudication committee’s charge is not defined. This reviewer would require that
`affirmative information must be supplied that the subject did not discontinue due to
`worsening disease. The presumption would be that all those who discontinue did so
`
`

`

`NDA 21-272 Remodulin ® (UT-15, Trcprostinil) ‘Pmtocol P0l:l3 rcvicvli
`
`04/18/02 1:09 PM page ll
`
`due to worsening of disease unless adequate data is there to be convincing that the
`status did not deteriorate.
`
`0 The sponsor should define which forms of hospitalization would not be considered as
`an event.
`
`0 The sponsor should collect data on the need for and adequacy of pain medication.
`
`APPEARS THIS WAY
`0N ORIGINAL
`
`
`
`

`

`
`
`This is a representation of an electronic record that was signed electronically and
`this page is the manifestation of the electronic signature.
`
`Abraham Karkowsky
`4/18/02 01:20:27 PM
`MEDICAL OFFICER
`
`

`

`Division of Cardio-Renal Drug Products
`Medical Officer Review
`
`NDA 2 l ~272
`
`Remodulin 7“ (treprostinol sodium) Injection
`Reviewer: Abraham M. Karkowsky, M.D., Ph.D.
`
`Date: July 13, 2001
`
`This document amends a statement in the review dated June 19, 2001. In that document this
`reviewer stated that no protocol was supplied for those Flolan patients transitioned to UT-lS (protocol
`mm 1). The sponsor (United Therapeutics) correctly pointed out that a protocol, amendments and CRFs
`were supplied with the original NDA. The conclusions of the June 19, 2001 document are not changed by
`this information.
`
`

`

`NDA 21-272
`
`Remodulin 7" (reprostinol sodium) Injection
`a
`
`06/1 2/011 1:09 AM
`
`page I
`
`Division of Cardio-Renal Drug Products
`Medical Officer Review
`
`NBA 21-272
`
`Remodulin T“ (treprostinol sodium) Injection
`
`This review covers the submissions dated 12 April 2001 and 15 May 2001 . This review
`incorporates Dr. Lawrence, the FDA statistician’s review for the April 12 submission. United
`Therapeutics submitted these two submissions in response to a meeting on 11 April 2001. At that
`meeting United Therapeutics proposed to analyze the combination of six-minute walk and the
`Borg-dyspnea score by combining these metrics into a single value. The rationale for combining
`these two parameters is they are the least likely (but not totally) devoid of bias due to unblinding
`by the nearly universal presence of infusion site pain limited to those treated with UT—l 5.
`Placebo-treated subjects mrely had infirsion site pain.
`
`United Therapeutics also submits additional data on the use of narcotic pain medication
`among those in treated longoterm with UT-l 5.
`
`Lastly, the sponsor submits information on a total of eight patients who were transitioned
`to UTJS after adverse events while receiving Flolan infusions.
`
`1) Additional analyses combining six-minute walk and Borg-dyspnea Index
`
`The sponsor's analysis was not pre-specified and only considered after the results of the
`study were available. The overall process of defining a metric was to combine normalized (to a
`scale of 0—1) rankings of six minute walk and normalized (to a scale of 0-1) ranking of the Borg—
`dyspnea scale. The resultant values were summed and then the subjects were then re-ranked with
`the ranks normalized (on a scale of 0-1). The sponsor graphs the % of patients in each group who
`achieve at least a given rank versus the rank scale that spanned the range of 0 to 1.
`
`The sponsor treats dropouts in different ways. Two extremes of these analyses are shown
`below. The first analysis (Figure 1) treated those who died, were transplanted, discontinued due
`to worsening of disease or who were too ill to perform the assessment as worse outcomes. All
`others who discontinued had last observations carried forward. Since there were far more
`
`discontinuations in among those treated with UT-lS than among those treated with placebo,
`additional analyses that assigned worse rank to subjects who discontinued based on criteria such
`as death or transplantation during the 100 days since randomization even after discontinuation
`due to adverse events. Other analyses also treated those who received flolan during the first
`month or those who received flolan during the 100 day span of the study as worst case scenarios.
`The analysis shown as Figure 2 treats all subjects who discontinued as worst outcomes
`independent of the reason for discontinuation and is the most conservative of the analyses
`performed by the sponsor. [Comment The most conservative analysis would treat he UT-15 as
`worst outcome and censor placebo patients]
`
`Based on these two analyses, there was a difference among those treated by placebo and
`those treated with UT-15. The interpretation of these curves however is obscure. Perhaps the only
`interpretation is that approximately 57-60% of those treated with UT-l 5 did better than the
`median (rank of 0.5) than the 4340% of those treated with placebo. Describing this benefit in
`words is not easy and would require demnvolufing the ranks. In essence all that can be said is
`that there is little benefit in walk distance (< 10 meters) and some benefit in Borg—dyspnea
`measurements.
`
`
`
`

`

`NDA 21-272
`
`Remodulin 7' (treprostinol sodium) Injection
`
`06/12/0l l 1:09 AM
`
`page 2
`
`This interpretation adds little to the information currently available.
`
`The very small pwalues associated with this analysis is hardly surprising. As Dr.
`Lawrence notes in his review
`I “ [W]hcn the p-value fi'om one variable is very small and this variable is combined with a
`second variable, it should not be surprising that the p-value from the sum of the two
`variables is small”.
`
`One last point, the Borg-dyspnea measurement is not entirely devoid of the problem of
`unblinding. The methodology that was employed by the sponsor was however, about a 5 good as
`it gets. There was a designated individual who not involved in the subject’s care elicited this
`metric. Unblinding, however, may have occurred at the level of the subject-investigator. The
`investigator likely knew the subject’s treatment based on the presence or absence of infusion site
`pain.
`
`Again combining the six-minute walk with the Borg index does not really add much to
`what we know a