CENTER FOR DRUG EVALUATION AND
`
`RESEARCH
`
`APPLICATION NUMBER: V
`
`20-986/SE3-003
`
`a
`
`MEDICAL REVIEW(S)
`
`

`

`DIVISION OF METABOLIC AND ENDOCRINE DRUG PRODUCTS (HFDoS I 0)
`
`MEDICAL OFFICER REVIEW
`
`APPLICATION #: #20986
`SPONSOR: NovoNordisk
`
`APPLICATION TYPE: NDA: Supplement
`PROPRIETARY NAME: NovoLog
`
`CATEGORY 0F DRUG: Diabetes
`Insulin analogue
`
`‘USAN I Established Name: x-14, Insulin aspart
`ROUTE: SQ infusion via
`
`1
`
`1
`
`MEDICAL REVIEWER: Elizabeth Koller
`
`REVIEW DATE: 12/20/01
`
`external pump
`
`SUBMISSIONS REVIEWED IN THIS DOCUMENT
`
`Submission Type: Comments:
`
`5 Document Date:
`‘
`'
`
`12120/00
`
`CDER Stamp
`Date:
`12121/00
`
`»
`
`, 2/7/01
`
`2/28/01
`? 3/6/01
`
`I 3/7/01
`‘ 3/22/01
`
`2/3/01
`
`2/28/01
`3/6/01
`
`3/7/01
`3/23/01
`
`SE3-003
`
`553-003 c
`
`553-003 IN
`SE3-003 IN
`
`SE3-003 IN
`N-000 0
`
`IND 1 -----~---—
`
`Comments on spreadsheet request
`
`
`W, 7 ~-
`Adverse event re orts su
`estin
`injection/infusipr‘i site (eagc ions,g
`InsulIn_InstabIlIty, and Infuswn set
`occlusIon
`
`13/29/01
`'
`'
`1 05/01
`
`r
`
`'
`
`5129111- r s.
`' T “
`1 0/5/01
`
`“55
`P‘OM
`
`1 10/24/01
`
`I 10/25/01
`‘ 12111101
`12/12/01
`12/17/01
`
`‘ 12/1 8/01
`12l19l01
`12/21/01
`
`10/2 5/01
`
`10/26/01
`12/13/01
`12/13/01
`12117101
`
`1 2/20/01
`12/20/01
`12/21/01
`
`SE3-003 BL
`
`SE3-003 BC
`SE3-003 BL
`SE3-003 BL
`SE3v003 IN
`
`SE3-003 C
`SE3-003 BL
`SE3-003 IN
`
`’
`
`Document Date:
`12/21/99
`Multiple
`Multiple
`
`RELATED APPLICATIONS (if applicable)
`
`APPLICATION Type:
`NDA #20563 SE3-024
`NDA #20563 Periodic reports
`
`Comments:
`Pump supplement
`Adverse events
`
`Overview of Application/Review:
`Three’, randomized, open-label, parallel-design studies of varying lengths were conducted in
`,
`the US. in adult patients with variable degrees of diabetic control. Pump studies comparing
`glycemic control (HgbAlc) using different types of insulin were conducted in IDDM patients 5
`‘ already familiar pump therapy. Studies in NIDDM patients new to intensive insulin therapy
`were conducted to compare glycemic control (HgbAlc) using X44 injection versus X—l4
`infusion therapy. The design of the clinical studies differed substantively from standard
`
`

`

`medical practice and typical use in real life, with more frequent changes of insulin, infusion
`sets, and infusion sites, Glycemic control and rates of hypoglycemia were comparable for the
`various treatment arms in all three studies The data from the smallest study (018; n=29)
`suggested that the time to infusion set failure/occlusion was shorter and the number of
`infusion set changes greater in the X-14 treatment arm than in the buffered human insulin‘
`arm. Although data on infusion set changes were also collected in the two, larger studies,
`these data were not provided. In addition, even though in vitro data indicated that X—l4
`consistently failed in the Mini-Med 506 pump on day 3, no correlation with the clinical data
`could be made because there were no records of the specific pumps used by individual
`patients.
`
`Outstanding lssues:.
`l-X-l4 can be approved for use in specific pumps with specific infusion sets. Approval
`cannot be extrapolated to other equipment. The X-l4 insulin, infusion sets, and infusion sites
`may be uséd for up to 48 hours.
`Z—Because the design of the clinical studies differed substantively from standard medical
`practice and typical use in real life, with more frequent changes of the insulin, infusion sets,
`and infusion sites, the consequences of such deviations from recommendations should be
`delineated in both the physician and patient labels. Because current standard practice reflects
`the large reservoir size, the software restrictions for reducing the amount of insulin inserted
`into the reservoir, and pump manufacturers’ printed instructions (including changing tubigg
`every threeWthephysieianandpafienfiabels must cTeafly?IE?catethat the directions for
`the specific use of X- 14in pumps supercede the manufacturers’ general guidelines.
`3-The sponsor should collect information on actual-use either in a phase four study in which
`insulin and infusion sets are not provided or via collection of adverse event data
`systematically using a questionnaire specifically designed‘to identify the causes of pump—
`insulin problems.
`
`4—New pump guidelines should be developed.
`
`Recommended Regulatory Action:
`
`New Clinical Studies:
`
`Clinical Hold
`
`Study May Proceed
`
`NDAs:
`
`Efficacy I Label Supp: with label chan
`
`s & a
`
`
`NotAlys __
`
`
`Date:
`Signed: Medical Rev-ewer: Eli
`Date:
`Medical Team Leader:
`
`
`
`oller
`
`.
`
`.
`
`f um insulin ma
`
`notion Approvable
`
`1212101
`
`APPEARS THIS WAY
`0N ORIGINAL
`
`

`

`1.-Medical Officer Review
`
`1.1 .-Administrative summary
`1.1.1.-NDA: #20986 SE3
`l.1.2.-Review: #1
`
`l.1.3.-Submissions:
`
`1.1.3.1.-Paper submission: 1 2/21/00
`1.] .3.2.-CANDA submission: none
`
`1.1.3.3.-Major amendment: none
`1.1.3.4.-Other submissions:
`2/7/01 SE3—003 C
`2/28/01 5133—003 IN
`3/6/01 SE3-003 [N
`3/7/01 8133-003 IN
`
`3/22/01'N—000 C comments on spread sheet request
`6/21/01 3133-003 BM Excel spread sheets
`8/29/01 SE3~003 IN
`
`10/5/01 P-004 adverse event reports suggesting skin reactions, insulin instability, & infusion set occlusion
`10/24/01 8133—00} BL
`10/25/01 SE3-003 BC
`12/11/01 8133-003 BL
`12/12/01 81-33-003 BL
`12/17/01 8133-003 IN
`
`JZJLMQLSB-OOlC..
`12/19/01 SE3-003 BL
`12/18/01 SE3-003 IN
`
`-
`
`,., ,9,
`
`-
`
`1.1.3.5.-Review completed: 12/21/01
`1.2.-Drug name
`1.2.1.-Generic name: insulin aspart
`l.2.2.-Trade name: NovoLog
`1 .3.-Sponsor: NovoNordisk
`1.4.—Pharmacologic category: diabetes, insulin analogue
`1.5.-Proposed indication: use in external pumps for subcutaneous infusion
`1.6.-Dosage form and route-of administration:
`1.6.1 .—Dosage form: vials for extraction of insulin that is to be put into a pump reservoir
`1.6.2.»Dosage: to be titrated using pre-prandial boluses and basah-ates ofContinuous
`infusion
`
`1.6.3.~Route-of-administration: subcutaneous infusion
`
`l.7.-NDA drug classification: standard
`1.8.-Imp0rtant related drugs: human insulin (semi-synthetic and recombinant)
`Lilly buffered human insulin, BR
`(approved; no longer marketed)
`NovoNordisk buffered human insulin, Velosulin
`lispro
`M
`
`/ A
`
`-
`
`1.9.- Related reviews: NDA #20563 pump reviews and adverse event reports
`W
`1.10.-Materials reviewed:
`
`1.10.1.-NDA #20986
`
`.._..._..... "HP—fl
`
`

`

`12/20/000 SE3—003 (38 volumes)
`2/7/01 8133-003 C
`2/28/01 8133—003 IN
`3/6/01 5133-003 IN
`.3/7/01 8133-003 1N
`
`3/22/01 NsOOO C comments on spread sheet request
`6/21/01 8133-003 BM Excel spread sheets
`8/29/01 5133-003 [N
`
`’
`
`10/5/01 P-004 adverse event reports suggesting skin reactions, insulin instability, & infusion set occlusion
`10/24/01 8133-003 BL
`”
`10/25/01 8133-003 BC
`12/11/01 SE3-003 BL
`12/12/01 8133-003 BL
`12/ 17/01 SE3—003 l'N
`12/18/01 SE3-003 C
`12/19/01 5133-003 BL
`12/18/Of 8133—003 IN
`
`1.10.2.-Other
`
`Draft pump guidance (1985) (appendix 1)
`Velosulin label
`
`Pump questionnaire developed by Dr. Koller in reSponse to questions raised by A.
`Morrison (Devices) and adverse event reports; 10/ 10/99
`lntemal e-mail regarding composition of tubing and needles for various infusion sets (P.
`Cricenti and V. Nakayama; 1 1/28/00)
`Mini-Med pump video and print information
`, s ~" ' r ’ ’ "“Disefionrc pump Video and print information
`Pump chat room
`1.10.3.-Safety update: none submitted
`1.] 1.-Table of contents
`1,—Administrative issues
`
`7
`
`r
`
`A
`
`,,
`
`_
`
`,
`
`7
`
`A7 J ,,
`
`,
`
`.
`
`a
`
`~
`
`21- Executive summary
`3.-lntroduction
`
`4.-Prior agreements
`5.—Objectives
`6.-CANDA
`7.-Financial disclosure
`W
`9.—Data integrity
`10,-Chemistry issues
`1 1.-Pharmacology-toxicology issues
`12.-Pharmacokinetic and pharmacodynamic issues
`l3.-Study design
`13.1.—General
`
`l3.2.—Demographic features and patient disposition
`13.3.-Efficacy variables
`13.4.-Safety variables
`14.-Efficacy results
`15.-Safety results
`16i-Reviewer’s commentary
`17.-Regulatory conclusions (including pump questionanaire)
`l8.-Label review
`l8. 1 .-General
`
`18.2.~Physician label
`18.3.-Patient label
`
`‘
`'"
`
`, ~
`
`3
`
`5
`7
`
`8
`9
`9
`9
`9
`9
`9
`10
`11
`11
`1 1
`
`12
`12
`13
`13
`13
`17
`l9
`21
`21
`
`22
`35
`
`

`

`.‘Executive summary
`The Diabetes Control and Complications Trial (DCCT) determined that good glycemic
`control decreased the risks for long-term complications in patients with Type 1 diabetes
`mellitus (IDDM). Intensive therapy was associated with lower HgbAlc values and better
`clinical outcomes than conventional therapy. Intensive therapy involves pre-prandial °
`dosing with a more rapid-acting insulin. The insulin is delivered either as a subcutaneous
`injection or a bolus infused subcutaneously by pump. Basal insulin needs are addressed
`with injections of longer-acting insulin or continuous low rates of insulin infusion.
`Although comparable glycemic control can be obtained with either multiple injections or
`pump infusion, some patients prefer pump therapy because of the ability to program low
`flow rates during sleep or exercise and because of the easy access to insulin for bolusing
`to manage both spontaneous and planned meals. Some patients are unable or unwilling to
`undertake the complexities of pump use and/or the associated costs of insulin
`administration. Initial pump costs are in theM Replacement of an ,
`infusion set costs approximately ~ . Furthermore, unlike intensive therapy with multiple
`injections, intensive therapy with pump infusion depends on the complex interaction of
`numerous variables including:
`a.—user skills.
`
`b.——accuracy, reliability, and other engineering features of the pumps.
`c.—reliability of the insulin in an atypical environment:
`i.-~non:reti'igerated andsexposedlowariabletandvperhaps'altematingjtemperatures
`determined, in part, by the ambient temperature, pump location e.g., axilla vs waist
`pocket, and heat absorption because of the color of the pump case or cover,
`iiHagitation that may precipitate drug aggregation,
`iii—exposed to variable lengths/areas of plastic and/or teflon cannula/tubing/insulin
`reservoir surfaces that may absorb drug/drug preservative, and kink or otherwise
`impair insulin flow, and
`iv.——exposed to low or absent flow rates because of programming or use of catheter-
`tubing devices that can be temporarily disconnected from the infusion site.
`d.--skin changes that may occur at the infusion site because of antigenicity of the insulin
`or the infusion set components.
`e.--altered residence times in the SQ tissue that may affect the risk for hyperglycemia and
`ketosis in the event of pump and/or insulin failure. Residence time may be a'property
`of the insulin itself or the skin changes induced by the insulin.
`
`Some of the complex interactions between pumps, infusion sets, and their insulins were
`recognized in the 19805 and delineated in a 1985 draft guidance. Additional in vitro
`' testing was required. Because it was known that inadequate buffering of human insulin
`resulted in frank occlusion, these guidelines emphasized buffer assessment. Only limited
`clinical testing was required because the chemistry parameters that impacted on the
`clinical safety of the drug had been identified. Since the development of these drafi
`guidelines, however, there have been many changes in the pumps, infusion sets, and
`insulin products. Because patients and physicians were anxious to utilize the increased
`rate of insulin absorption to permit more rapid adjustment of blood glucose, rapid—acting
`analogues were used offclabel in pumps. Data from lispro, the first rapid-acting insulin
`analogue, however, suggest that the clinical and in vitro guidelines developed for human
`
`

`

`insulin in pumps are not sufficient for insulin analogues, e.g. infusion site reactions
`(pump bumps) are not uncommon, and, e.g., temperature changes alter pH and the
`stability.
`
`In the current submission, the sponsor provided data on Mini-Med 506 and Disetronic H~
`TRON pump-insulin function at 37°C and with continuous agitation. The durability and
`performance of X-14 in the Disetronic pump exceeded that in the Mini-Med pump which
`repeatedly became occluded on day three and thereafter. The sponsor did not provide
`information on pump-insulin function afier exposures higher than 37°C. The sponsor also
`did not provide information on insulin stability afier temperature cycling, These types of
`exposures are likely in View of the appliances and actual-use videos that the pump
`manufacturers provide. Although glycemic control, as measured by HgbAlc, was
`comparable for buffered human insulin and X-l4, the clinical data were collected under
`optimal conditions. Patients changed insulin, infusion sets (reservoirs, tubing, and
`catheters), and infusion sites at intervals that did not exceed 48 hours. Patients were given
`as much insulin and as many infusions sets as needed. The sponsor did not conduct any
`studies that approached real-life use-owith insulin, infusion set, and site changes every
`three to seven days. The data from the smallest study (018; n=29) suggest that the time to
`infusion set failure/occlusion was shorter and the number of infusion set changes greater
`in the X 14 treatment arm than in the buffered human insulin arm. Although data on
`
`7 infusion set “changes werealsacolleetedsrrthfiwe-largerstud'res, these data Were net
`provided No correlation of the in vitro pump function data with the clinical data could be
`made because there were no records of the specific pumps used by individual patients.
`
`Changing infusion sets and sites are not just cost and convenience issues They are safety
`issues Any insulin instability or predilection for occlusion (whether directly by
`precipitation or indirectly by infusion site reaction) can resultin more hyperglycemia and
`diabetic ketoacidosis. The rapid onset of hyperglycemia and diabetic ketoacidosis
`observed with the shorbacting insulin analogues is likely to be accentuated with pump
`use because the insulin reservoir in the skin is smaller than with injections. Any
`progressive insulin instability may also result in continual upward dose titration
`complicated by hypoglycemia when the reservoir is refilled with new insulin. Patients
`who are switched from buffered regular insulin in pumps and patients who are switched
`from multiple«dose intensive injection therapy with either human insulin or rapid acting
`analogues may be at particular risk for developing these problems.
`
`The sponsor provided incomplete information in this submission, but there is probably
`sufficient information to write a narrow label that will provide adequate patient
`protection. This is an important consideration because the drug is likely to be used off-
`label, and it would be desirable to avoid the morbidity and mortality observed with lispro.
`The sponsor should collect information on actual—use either with a phase four study in
`which insulin and infusion sets are not provided or via collection of adverse event data
`systematically using a questionnaire specifically designed to identify the causes of pump-
`insulin problems
`
`,meMWN ..
`
`

`

`3—Introduction
`
`3.1.~—Rationale for diabetes management with pump therapy
`The Diabetes Control and Complications Trial (DCCT) determined that good glycemic
`control decreased the risks for long—term complications in patients with Type 1 diabetes
`mellitus. Intensive therapy was associated with lower HgbAlc values and better clinical
`outcomes than conventional therapy. Typically, intensive therapy involves pre-prandial
`dosing with a more rapid-acting insulin in conjunction with a longer agting insulin to
`provide a basal level of control throughout the day. Four or more injEctions daily are
`required. Alternatively, patients can utilize subcutaneous insulin infusions delivered by a
`pump. A basal rate is based on the anticipated activity level. Insulin boluses are given to
`cover food consumption. Additional insulin is given in the event of unexpected
`hyperglycemia. Conversely, insulin rates are reduced in the event of hyperglycemia.
`
`Intensive therapy, whether multiple injection or by pump infusion, requires frequent
`monitoring of blood glucose. Fingerstick (forearm) sampling is typically performed
`between four and six times per day. Some patients are unable or unwilling to undertake or
`continue intensive therapy because of the number of insulin injections, the complexities
`of pump use, the fiequency of glucose monitoring, and/or the associated costs of
`increased glucose sampling and insulin administration. Replacement of an infusion set
`costs approximately .- The “tight control” that can be achieved with intensive therapy
`,isalsoassociated-withanénereasedfiskofhypogiytemia.TfieTFeque'néy’6r’sév’enfyflof I»
`hypoglycemia may be the limiting factor to HgbAl c reduction-either directly through
`symptoms or indirectly through the subsequent glucose rebound. Although comparable
`glycemic control can be obtained with either multiple injections or pump infusion, some
`patients prefer pump therapy because of the ability to program low flow rates during
`sleep or exercise and because of the easy access to insulin for bolusing to manage
`spontaneous or planned meals.
`
`3.2.——Issues in the development of insulins for external pumps
`Unlike intensive therapy with multiple injections, intensive therapy with pump infusion
`depends on the complex interaction of numerous variables including:
`a.——user skills.
`-.-
`
`b.——accuracy, reliability, and other engineering features of the pumps.
`c.—reliability of the insulin in an atypical environment:
`i.-—non-refn'gerated and exposed to variable (and perhaps alternating) temperatures
`determined, in part, by the ambient temperature, pump placement e.g., axilla vs waist
`pocket, and heat absorption because of the color of the pump case or cover,
`ii—agitation that may precipitate drug aggregation,
`iii—exposed to variable lengths/areas of plastic and/or teflon cannula/tubing/insulin
`reservoir surfaces that may absorb drug/drug preservative, and kink or otherwise
`’impair insulin flow, and
`iv.—exposed to low or absent flow rates because of programming or use of catheter-
`tubing devices that can be temporarily disconnected from the infusion site.
`d.--skin changes that may occur at the infusion site because of antigenicity of the insulin
`or the infusion set components, e.g., nickel, or infection that may occur because of
`extended insulin storage without refrigeration and with alterations in bacteriostatic
`
`

`

`additives.
`
`e.--altered residence times in the SQ tissue that may affect the risk for hyperglycemia and
`ketosis in the event of pump and/or insulin failure. Residence time may be a property
`of the insulin itself or the skin changes induced by the insulin.
`
`Some of the complex interactions between pumps, infusion sets, and their insulins were
`recognized in the 19805 and delineated in a 1985 draft guidance. Additional in vitro
`testing was required. Because it was known that inadequate bufferingaf human insulin
`resulted in frank occlusion, the guidelines emphasized buffer assessmefit. Only limited
`clinical testing was required because the chemistry parameters that impacted on the
`clinical safety of the drug had been identified. Since the development of these draft
`guidelines, however, there have been many changes in the pumps, infusion sets, and
`insulin products. Some pumps operate on a time basis, e.g, Disetronic pumps every three
`minutes. Others operate on a volume basis, e.g., Mini—Med pumps in 1/10 insulin unit
`increments. Tubing can now be over 100 cm to facilitate distal placement. Appliances
`permit pumps to be used in the sauna or shower. Data from another insulin analogue
`suggests that temperature changes alter pH and the stability of the insulin. Given this
`proliferation of change, it is quite possible that the in vitro criteria used to assess buffered
`human insulins are not adequate predict problems that could arise with insulin analogues
`in the clinical setting. Furthermore, because pump therapy15 so equipment intensive, it is
`possible that clinical trials will not identifyproblemsjhaLcouldansarmthesettmgof-lessw
`superv151on or less replacement equipment/insulin.
`
`3.3.—Rationale for the development of X-14 (insulin aspart; NovoLog) in pumps
`_ X-l4 is a modified insulin. The substitution of asparticacid for proline in position 28 of
`the B—chain reduces aggregation of the insulin molecule with other insulin molecules and"
`facilitates the absorption of the modified insulin through the skin. This pharmacokinetic
`property ismost evident when compared to more concentrated human insulin (U100 and
`U500). It is less evident with less concentrated human insulin (U40). Because the volume
`of infused insulin is relatively small at any given time, the insulin concentration in the
`skin is low, and the phannacokinetic differences between X-14 and-human insulin could
`be expected to be less pronounced in pump patients than in injection patients.
`Nonetheless, the sponsor believes that the phannacokinetic—phamr’acodynamic (PK-PD)
`profiles are sufficiently distinct that patients will find convenience in bolus dosing
`immediately prior to meals.
`
`4.—Prior agreements
`4.1.———Drafi guidelines
`Draft guidelines were developed in 1985 prior to the development of insulin analogues.
`(See appendix 1.)
`
`4.2—Other
`In a three telecommunications and two letters between 3/17/99 and 4/27/99, the sponsor
`was informed that the FDA did not have guidelines in addition to those proposed in 1985,
`but that some clinical data suggested that the proposed guidelines might not be sufficient
`for insulin analogues. Potential stability problems with heat were discussed with Dr. Poul
`
`

`

`Strange. Devices that could be used during the clinical trials to assess of maximum
`temperature exposure was also discussed. Assessment of the pump in real life setting was
`encouraged because it is known that financial constraints limit changes in tubing and
`_ reservoirs to three to seven day intervals. The sponsor attempted to address this by having
`patients use a single infusion set during the seventh and final week of a small study (018).
`
`The sponsor did not request a pre-NDA meeting. No advance notice of the NDA-
`‘5';
`supplement was provided prior to its arrival.
`3"
`at
`
`5.——Objectives
`The sponsor has sought to show that the insulin analogue, X-14, can be used in external
`pumps for continuous subcutaneous administration without additional modification of the
`drug substance or excipients.
`O
`
`6.—CANDA
`
`There was no CANDA submission. Supplement data were provided on EXCEL spread
`sheets 6/21/01.
`
`7.—-Financial disclosure
`
`The sponsor stated that there were no financial interests to disclose (Volume 1).
`
`’8.'—:Pea'iaiii’c i‘v’ai'v’éf ’
`
`No pediatric waiver was requested.
`
`9.—Data integrity
`Most of the investigators were boarded in endocrinology-although some were pediatric "
`endocrinologists. The curriculum vitae were short or truncated-limiting assessment of
`qualifications. The investigator examining the insulin for crystals and pH changes was
`not a chemist. The DSI inspectibns at two sites :W
`V) were reportedly unremarkable.
`‘
`
`10.——Chemistry issues
`Recombinant X-l4 insulin analogue has the empirical formula C257H331N65079$6 and a
`molecular weight of 5825.8. Each milliliter contains X-l4 (100 units), glycerin (16
`mg/ml), phenol (1.50 mg/ml), m—cresol (1.72 mg/ml), zinc 19.6 (ug/ml), disodium
`hydrogen phosphate dihydrate (1.25 mg/ml), sodium chloride (0.58 mg/dl). The pH is 7.2
`to 7.6. The insulin was previously approved for use as an injectable. The insulin was not
`- modified for the proposed pump indication. The size of the pump reservoirs is 3 ml. The
`sofiware in some pumps encourages patients to fill the reservoir completely.
`
`The, chemistry testing was most notable for the decrease in the cresol, which serves as a
`preservative and anti-microbial agent, and the pump stoppages with the Mini-Med
`equipment. (See table 1.) The stoppages were recurrent on day three and thereafter. There
`were also increases in ‘BZSisz—l4 and other X-l4 related substances over time so that
`
`some samples did not meet release limits at six days.
`
`

`

`Table 1 Chemistry Testing for X-l4 as 3 Pump Insulin
`Potency, Purity, Degradants Disetronic Pump (H—tron)
`Classic tubing 80 cm
`Tenders tubing 80 cm
`3 insulin lots
`
`
`
`0.9 U/hr + 3 6U boluses/day
`0.9 U/hr + 3 6U boluses/day
`
`
`At 37°C
`At 37°C
`
`
`
`Appearance
`
`Tubing change at days 2 & 6
`Agitation throughout study
`Increase in particles in elute &
`especially the syringe. Meets
`USP.
`
`Mini-Med Pump (506)
`_
`Polyfin tubing- 106 cm
`Sof-set Polyfin tubing 106 cm
`3 insulin lots
`.
`
`
`Tubing chgige at days 2 & 6
`
`. Agitation-throughout study
`
`Increase in particles in elute &
`especially the syringe. Meets
`USP.
`
`
`
`Volume
`
`~----—
`
`——-—~_._
`
`,
`
`,
`
`pH
`
`No leakage reported
`No leakage reported
`.
`Increase 0.05 U by day l
`Increase 0.07 U by day 1
`
`
`
`No further increase dayl to 6
`No further increase dayl to 6
`-
`. .w
`Range
`Range’, N
`
`
`pH delta does not —>particulates
`pH delta does not ->particulates
`
`
`
`
`No sterility assessment
`Microbial Growth
`No sterility assessment
`M-cresol above limits at 2
`(Tested using pooled sampled M-cresol decreased over 4
`
`
`
`{mm agi'a'ad Syringes)
`days, but still above limits at
`days, below limits at 7 days
`
`
`
`
`
`) , {—Lea; " ' ‘ 'wn‘. iif!Ydaxs‘wrv I *'¥)#w_
`
`
`
`
`
`
`
`
`infusion system is permitted.)
`i
`Degradants
`
`
`
`
`Leachables
`
`’
`
`‘
`
`Pump Function
`
`No pump stoppages.
`
`Temperatures >37'C
`Thermocycling
`
`-
`
`Very Low Flow rates
`Use with Diluent
`
`Pump stoppages with bolus
`dosing on day 3 with both
`catheter types.
`Once stoppages occurred
`they could not be attenuated
`Not done
`Not done
`
`Not done
`Not done
`
`I 1.—Pharmacology-toxicology issues
`Alterations in the infusion sites have been observed in pigs (which have the skin most
`similar to that of human beings) with another insulin analogue. No animal data were
`submitted for this application.
`
`

`

`l2.—-—Pharmacokinetic-pharmacodynamic issues
`No PK-PD studies were submitted. The sponsor did not assess potential PK-PD changes
`that could occur in conjunction with infusion site changes.
`
`l3.-Study design
`l3.1.-General
`
`,
`
`-
`
`Three randomized, open-label, parallel—design studies of varying lengths were conducted
`in the US. in adult patients with variable degrees of diabetic controLg‘ables 2 and 3).
`Pump studies using different types if insulin were conducted in IDDM patients already
`familiar pump therapy. Studies in NIDDM patients new to intensive insulin therapy were
`compared X-l4 injection versus X-l4 infusion therapy (Table 3). Patients used a variety
`of pumps and infusion sets (Table 4). Records of the types of pump equipment were not
`maintained. Patients were to change their insulin, infusion sets (reservoir, tubing,
`cathetér) and infusion sites every 3 days at a minimum. The number of infiision set
`changes and episodes of hyperglycemia and DKA particularly unexpected hyperglycemia
`along with the number of episodes of hypoglycemia. Infusion sites were inspected
`weekly for study 018 and every 2 to 4 weeks for studies 023 and 024. The insulin and
`infusion sets were visually inspected in study 018. Although infusion sets were changed
`every 48 hours or sooner, in the final week of study 018, infusion sets were retained until
`failure. Various parameters of glycemic control were also assessed.
`
`’ ” "’“I‘atflifl 7555126 Fathl‘es 0f Studies
`
`Diabetes
`DrugsDelivery
`Design
`E Type l
`Velosulins—Pumpr
`Parallel
`
`X- I4—Pump
`Parallel
`X—l4~—Pump
`Type 2
`-- X‘W‘Wemo" -
`Velosulin--Pump
`Parallel
`
`023
`
`Blinding
`No
`
`N0
`
`N
`
`
`
`X-l4--PumpI isismsr =
`
`l Velosulin (V) dosed 30 minutes before meals; Lispro (LP) and X-l4 dosed 515 minutes before meals.
`2 NPH used as basal insulin once or twice daily.
`’ :
`3 The 7'h week was used to assess the time interval an infusion set could be used before obstruction or hyperglycemia
`4 Patients were familiarized with more intensive therapy over 4 weeks. An 8 week dose titration phase
`post—randomization preceded the 16 week treatment phase.
`"
`
`?
`
`Yes—but not naive to insulin use-)
`
`Entry HgbAlc
`(% units)
`
`Entry Age
`(yrs)
`
`37; <12
`
`335
`
`218
`36,5; <9
`No-use for 33 mo;
`
`*CSll=continuous subcutaneous insulin infusion=pump.
`
`Table 3 Selected Entry Criteria
`Naive to X- l4 Naive to CSll‘
`
`
`
`
`024,
`
`?
`
`
`
`
`
`
`ll
`
`
`
`

`

`Table 4 Pump Equipment Parameters
`Study
`Pump Type
`Infusion
`
`
`
`Tubing
`
`
`
`Infusion Site
`
`018
`
`023
`
`Mini-Med
`
`Mini-Med
`
`?
`
`’?
`
`.
`
`.
`
`?
`
`?
`
`7
`
`?
`
`?
`
`?
`
`?
`
`?
`
`548 hrs
`
`le4 428
`
`ql wk
`
`548 hrs
`
`Collected;
`
`q2 wks x2 mo—>
`
`
`
`
`
`
`--I I ----_
`
`
`'
`
`
`
`‘A=axilla; B=breast; Wiwaist or hip; O=other.
`
`
`
`
`--I‘ I"
`
`only Disetronic)
`
`13.2.-Demographic features and patient disposition
`No special patient populations were studied. The patients with NIDDM were older than
`the patients with IDDM (Table 5). They also had higher HgbAlc values, which may
`reflect their prior treatment-which was conventional therapy, not intensive insulin
`therapy. There were different proportions of male and female patients in the different
`studies and sometimes between treatment groups within a study. There were similar
`numbers of patients in each treatment group who discontinued. The reasons for
`discontinuation were similar for the various treatment groups. Curiously, the number of
`days in the study was highly variable. Multiple patients continued in studies several
`weeks after the scheduled termination.
`
`
`
`Table 5 Demographic Features and Patient Disposition of Randomized Patients
`HgbAlc
`# DCed
`Days in Study
`(% units; n)
`(mean; range)
`50.0 (49-51)
`_ 48.3 (15-55) "
`7
`164.2 (39-204)
`162.5(1-217)
`_ 1062—11441)
`- 113.9 (32-180)
`121.1 (78-177)
`
`
`
`_ —_
`
`5 1
`
`l DC=discontinued; did not have a HgbAlc value at the final study visit.
`2 versus X-l4; p=0. l 2.
`3 versus X-l4; p=0. l3.
`C=Caucasian; B=Black; A=Asian; O=other.
`F=female; M=male.
`
`l3.3.—Efficacy variables
`The primary efficacy variables were HgbAlc and changes in HgbAlc. Secondary
`variables depended on self—collected glucose measurements at 8 points during the day.
`Mean glucose values at various times of the day were determined as were glucose
`excursions, mean glucose profiles, and variation from the mean glucose. Fructosamine
`values were assessed in study 018-primarily to determine if there was a progressive
`change in glycemic control during the final week of the study when the insulin, infusion
`set, and infusion site were not changed until failure.
`
`12
`
`w
`
`_,_..«,.~.... « _ —
`
`

`

`13.4.-Safety variables
`.
`Safety variables included the number of infusion set Changes and episodes of
`hyperglycemia and DKA (particularly unexpected hyperglycemia) along with the number
`of episodes of hypoglycemia. The insulin and infusion sets were visually inspected for
`precipitation and the insulin pH tested in study 018. The time to infusion set failurewas
`monitored in study 018. Injection or infusion site reactions were also monitored.
`
`!
`l4.-Efficacy results
`Glycemic control as measured by HgbAlc did not change substantially in IDDM patients
`treated with pumps regardless of the insulin used (Table 6). Glycemic control as
`measured by HgbAlc improved in NIDDM patients switched to intensive therapy, but
`did not differ by mode of delivery (injection versus pump). The Self-collected glucose
`measurements, particularly the change in mean glucose profile, correlated poorly with the
`more rigorous HgbAlc parameters so no further assessments of the self—collected values
`were performed.
`
`Table 6 Glucodynamic Assessments
`Study
`Drug
`HgbA 1c
`(% units; n)
`
`
`
`
`
`Glucose Profile
`(mg/dl; n)*
`
`
`
`
`Correlation (r; n) '
`HgbAlc vs Profile
`
`
`
`_——-————
`
`————-———
`024
`
`
`
`
`
`
`_ X-14-o.oo4(57>
`——————_—
`
`
`
`“ Serial, glucometer values self-collected over the course of a day with the collection days determined by the patient;
`all patients did not take 100% of measurements at all timepoints
`
`
`
`15.-Safety results
`lS.l.-Clinical trials
`
`Hypoglycemia did not differ between treatment groups, and hypoglycemia was not
`clearly correlated with HgbAlc or changes in HgbAlc (Table 7). Major increases in total
`insulin doses were not required to obtain comparable glycemic control with X-l 4
`compared to Velosulin (Table 8). Episodes of hyperglycemia and DKA were not more
`common in any particular treatment group. Because these episodes of hyperglycemia did
`not correlate well with HgbAlc, they were likely very transient (Table 9). Curiously,
`there were more hyperglycemic episodes per patient in the study with the lowest HgbAlc
`levels (018). Data from study 018 suggest that more infusion sets changes were required
`to maintain glycemic control in patients who used X-14 350.05). Unfortunately,
`although comparable data were collected in the subsequent studies, they were not
`presented (Table 10). The time-to-infusion set failure during the seventh week of study
`0187was also shorter for patients who used X-l4 (p50.05) (Table 10). The numbers of
`“unplanned infusion set changes for unexpected hyperglycemia” could not be interpreted
`because of the subjective nature of the determination. There were no clear