Division of Pharmaceutical Evaluation-II

Office of Clinical Pharmacology and Biopharmaceutics

NDA: 21-332 Relevant IND: 39,897

Brand Name: Symlin™ Generic Name: Pramlintide Acetate

Concentrations: Vial – 0.6 mcg/mL sterile injection in 5 mL vials

Cartridge – 1.0 mcg/mL sterile injection in 1.5 mL cartridges

Sponsor: Amylin Pharmaceuticals, Inc.

9373 Towne Centre Drive, San Diego, CA 92121

Submission Date: 7-DEC-2000 Division Due Date: 6-JUN-2001

5-APR-2001

Advisory Committee: 26-JUL-2001 PDUFA Date: 7-OCT-2001

CPB Reviewer: Steven B. Johnson, Pharm.D.

CPB Team Leader: Hae-Young Ahn, Ph.D.

Acknowledgements: Daniel Davis, M.D.; S.W. Johnny Lau, Ph.D.; Todd Sahlroot, Ph.D.

EXECUTIVE SUMMARY

On December 7, 2000, Amylin Pharmaceuticals submitted NDA 21-332 in support of Symlin™ (pramlintide acetate) injection. Pramlintide is the synthetic analogue of the 37-amino acid polypetide, amylin. The proposed mechanism of pramlintide action is complex, with regulation of postprandial glucagon concentrations and altered gastric emptying rate being the most well described. Two formulations of Symlin™ have been proposed for marketing, a 0.6 mg/mL (vial) formulation that will be administered by syringe and a 1.0 mg/mL (cartridge) formulation that will use a “pen” system for administration. Symlin™ has been proposed for use as adjunctive therapy to insulin in patients with type 1 or insulin-requiring type 2 diabetes mellitus (DM).

Included in this application were 28 clinical pharmacology and biopharmaceutics related studies or reports. Of these studies, 19 were used to make the CPB recommendation. Many of the early studies were found to be unacceptable for review due to formulation and/or assay issues. Common to the studies that were utilized in this review included: a formulation pH of 4.0 and/or the use of the current immunoenzymetric assay (IEMA) for pramlintide pharmacokinetic (PK) studies.

The following is a brief description of Symlin™ attributes. First of all, there is a high degree of inter-subject variability for all PK parameters, except t_½ and T_max. This drug is absorption rate limited, has a time to maximum pramlintide plasma concentration (T_max) of approximately 20 minutes, and a half-life of about 50 minutes. Pramlintide is metabolized to des-lysine pramlintide, which has 100% of the activity as pramlintide, and other none reactive fragments. There is no apparent drug accumulation following multiple doses in either type 1 or type 2 diabetes patients.

In order to take full advantage of pramlintide’s delayed gastric emptying effect, Symlin™ should be administered about 15 minutes before a meal – this timing would correspond with the pramlintide T_max. Symlin™ should be administered subcutaneously into the tissue of the anterior abdominal wall only, with a maximum dose of 360 mcg/day – divided BID, TID, or QID.

During the course of this review, a series of questions were generated to address pertinent issues that were thought to be key for the approval of this application. The most prominent of these questions are:

1) Is the analytical method used to detect pramlintide in human plasma precise and accurate?

Yes, the immunoenzymetric assay used to detect human plasma pramlintide exhibits precision and accuracy estimates that are acceptable. However, it should be noted that the samples used in the quality control analysis were sufficiently far enough away from the calibration limits and the lower limit of quantitation as to create some concern about the plasma concentrations that fall between the LLOQ and the lowest quality control sample.

2) Is there any assay interference from endogenous substances or metabolites?

Yes, this assay is susceptible to interference by endogenous amylin, the des-lysine pramlintide metabolite, and human anti-mouse antibody (HAMA). Also, the values reported as pramlintide concentrations are actually pramlintide, the pramlintide metabolite, and amylin.

3) What is the absolute bioavailability of Symlin™?

The bioavailability of a subcutaneously administered dose of Symlin™, relative to an equivalent intravenously administered dose of Symlin™, is approximately 37%. Pramlintide exhibits absorption rate limited pharmacokinetics.

4) What effect does pH, mixing, volume, or concentration have on the bioavailability of Symlin?

The formulation pH has a significant effect on the bioavailability of Symlin™. A formulation pH of 4.7 was shown to exhibit a 25% reduction in bioavailability compared to the to-be-marketed pH 4.0 formulation.

Compatibility studies, that would describe what substances could be mixed with pramlintide, were not performed. However, in interaction studies where pramlintide and insulins were mixed in the same syringe, pramlintide and sometimes insulin pharmacokinetics were significantly altered.

Volume and concentration had no apparent effect on Symlin™ bioavailability.

5) Does Symlin™ exhibit dose proportionality over the entire proposed dosing range, 30 mcg to 180 mcg?

No, Symlin™ does not exhibit dose proportionality over the entire proposed dosing range. However, studies in healthy volunteers and patients with type 1 and type 2 diabetes demonstrate a dose-related exposure. Apparent clearance in normal healthy subjects ranged from 2.36 L/min to 2.87 L/min for the dosing range of 30 mcg to 120 mcg.

6) Given that Symlin™ will be available in two concentrations, are these formulations bioequivalent?

Bioequivalence was established between the cartridge formulation administered by a pen system and the vial formulation administered with a syringe. Ninety-five percent confidence limits were within the 80% to 125% boundaries for both C_max and AUC.

7) Since there will be three suppliers of pramlintide material for Symlin™, are there any PK-related concerns about using multiple sources of this protein?

No, there are no outstanding concerns related to the multiple sources of pramlintide. Information submitted in the form of a BE study was sufficient to conclude that pramlintide material from UCB-Bioproducts and Bachem were equivalent. Material from both Bachem and Mallinckrodt was used throughout the development program and was not considered different.

8) Are there any differences between the pramlintide PK profiles of type 1 and type 2 diabetes patients?

Yes, there is a difference between the pramlintide PK profiles of type 1 and type 2 diabetes patients. In the studies evaluated in this application, patients with type 2 diabetes exhibited lower relative pramlintide plasma concentrations than did patients with type 1 diabetes. This fact is reflected in the proposed diabetes type-specific dosing regimens.

9) Pramlintide is renally eliminated, what effect does renal insufficiency have on pramlintide PK?

The exact effect that renal insufficiency has on pramlintide PK cannot be stated with absolute certainty due to the design limitations of the renal study (e.g., parallel design with 3 to 8 subjects per group and inherent inter-subject variability). However, there was no observed trend in the data to suggest that renal insufficiency adversely effects pramlintide PK. It should be kept in mind that pramlintide dosing in humans has spanned 30 mcg to 10 mg in single dose studies.

10) Can pramlintide be mixed with insulin(s)?

No, pramlintide should not be mixed in the same syringe with any insulin. The sponsor evaluated mixing effects of pramlintide with rapid, short, intermediate, and long-acting insulins. The results suggest that pramlintide PK, and sometimes insulin PK, is compromised when these agents are mixed.

11) Since pramlintide delays gastric emptying time, what effect does Symlin™ have on orally administered medications?

Two studies evaluated the effect Symlin™ has on the PK of Lo/Ovral, an oral contraceptive, and ampicillin, a relatively acid-stable antibiotic. Results indicate that norgestrel, the progestin component of Lo/Ovral, achieves significantly lower and delayed C_max values when administered 15 minutes after a dose of Symlin™. However, there was no difference in the extent of norgestrel exposure.

Conversely, pramlintide appeared to have no effect on ampicillin AUC or C_max. However, T_max was increased by approximately one hour.

The conclusion drawn from these studies would suggest that orally administered drugs that are expected to have a rapid onset of action, or those that are adversely affected (e.g., degraded) by prolonged gastric retention times, should be administered at least one hour prior to dosing pramlintide.

12) What effect does a morning dose of pramlintide have on the gastric emptying of a lunchtime meal?

There was no observed interference of a morning dose of pramlintide on a lunchtime meal eaten approximately 4 hours after pramlintide administration. This is consistent with pramlintide’s attributes of a relatively short half-life of pramlintide, ~ 50 minutes, and the fact that there is no indication of pramlintide accumulation.

Based upon the results of the studies that answer the above questions, it has been concluded that Amylin, Inc. has supplied sufficient information to the PK section for Symlin™, NDA 21-332, to have been adequately evaluated.

RECOMMENDATION

The Office of Clinical Pharmacology and Biopharmaceutics has reviewed NDA 21-292 for Symlin™ (pramlintide acetate) and finds the application acceptable, pending the indicated labeling changes (see Comments to Sponsor and Labeling Changes).

EXECUTIVE SUMMARY.................................................................................................................. 1

RECOMMENDATION....................................................................................................................... 3

TABLE of CONTENTS...................................................................................................................... 3

APPENDIX INDEX............................................................................................................................ 4

BACKGROUND – (from sponsor)...................................................................................................... 4

TERMS and ABBREVIATIONS......................................................................................................... 5

DRUG CHARACTERISTICS.............................................................................................................. 5

ANALYTICAL.................................................................................................................................. 6

HUMAN PK – BIOAVAILABILITY/BIOEQUIVALENCE......................................................................... 7

HUMAN PK – BIOEQUIVALENCE................................................................................................... 10

HUMAN PK – TARGET POPULATION............................................................................................. 12

HUMAN PK – PLASMA/BLOOD..................................................................................................... 17

HUMAN PK – EX VIVO – Placental Transfer..................................................................................... 17

PHARMACODYNAMICS................................................................................................................ 17

LABELING.................................................................................................................................... 19

COMMENTS TO THE SPONSOR.................................................................................................... 19

APPENDIX INDEX

Study #	Title	Page #
137-125	An open-label, randomized, four-period cross-over study in normal volunteers of the bioavailability of selected concentrations of pramlintide in two different formulations.
137-142	An open-label, randomized, two-period crossover study in healthy volunteers of the bioequivalence of two different formulations and dosage forms of pramlintide.
137-126	An open-label, randomized, four-period cross-over study of the proportionality of four subcutaneous doses of pramlintide (AC137) administered at a constant volume in normal volunteers.
137-127	An open-label, single-dose, pharmacokinetic study of pramlintide in type 1 diabetics with renal impairment.
137-133	A randomized, double-blind, placebo-controlled, single-dose, two-period cross-over study to evaluate the effect of pramlintide on the pharmacokinetics of ethinyl estradiol and norgestrel in healthy female subjects receiving the oral contraceptive ageng Lo/Ovral^®.
137-134	A randomized, double-blind, placebo-controlled, single-dose, two-period cross-over study to determine the effect of pramlintide on the pharmacokinetics of ampicillin in healthy subjects.
137-130	A randomized, double-blind, single-dose, two-period cross-over study of the safety of pramlintide and lispro insulin administered as two separate subcutaneous inject5ions in conjunction with NPH, Lente, or Ultralente insulin in patients with type 1 diabetes mellitus.
137-115	An open-label, randomized, cross-over study in type 1 diabetes mellitus of the pharmacokinetics of subcutaneous pramlintide (AC137) and 70/30 insulin mixed together and as separate single injections.
137-119	An open-label, randomized, five-period cross-over study of the pharmacokinetics of subcutaneous pramlintide (AC137) versus placebo plus NPH and regular insulin mixed together and as separate single injections in patients with type 1 diabetes mellitus.
137-145	An open-label, randomized, two-period cross-over study in healthy volunteers to test the bioequivalence of pramlintide supplied by two different manufacturers.
137-143	An open-label assessment of the single dose and multiple dose pharmacokinetic profiles of pramlintide in subjects with type 1 diabetes mellitus.
137-144	An open-label assessment of the single dose and multiple dose pharmacokinetic profiles of pramlintide in subjects with type 2 diabetes mellitus.
137-120	An open-label, randomized, five-period cross-over study of the pharmacokinetics of subcutaneous pramlintide (AC137) versus placebo plus isophane and soluble insulin mixed together and as separate single injections in patients with type 1 diabetes mellitus.
137-118	The effect of single doses of pramlintide on the gastric emptying of two meals.
137-137	A study to determine the effect of pramlintide on the gastric emptying of subjects with type 2 diabetes mellitus currently requiring insulin.
REST 98049	Characterization of pramlintide metabolites following bolus subcutaneous administration in type 1 diabetic subjects.
REST 98044	Characterization of binding of pramlintide to components from rat, dog, rabbit, mouse, and human blood.

BACKGROUND – (from sponsor)

In people without diabetes, plasma glucose concentrations are tightly regulated by the co-secretion of the hormones amylin and insulin from the pancreatic b-cells in response to nutrient intake, and by the glucagon secretion from pancreatic a-cells in response to a variety of stimuli, including hypoglycemia and elevated concentrations of amino acids. In people with type 1 diabetes, the pancreatic b-cells are usually destroyed by an autoimmune process, leaving patients deficient in both insulin and amylin. In people with type 2 diabetes, insulin resistance leads to an increased demand for insulin, and initially results in increases in b-cell secretion of both insulin and amylin. Over time, b-cell secretion fails, and relative deficiencies of both insulin and amylin occur in conjunction with inappropriate fluctuations in glucose concentrations, overt hyperglycemia, and increased risk of hypoglycemia.

Results of nonclinical and clinical studies indicate that the 37-amino acid polypeptide amylin and the amylin analogue pramlintide contribute to glucose regulation through several mechanisms including reducing the postpranidal rise in glucagon concentrations without impeding the glucagon response to insulin-induced hypoglycemia, and regulating the rate of nutrient delivery to the small intestine via an effect on gastric emptying. It has been proposed that amylin’s effect on gastric emptying may be exerted via a central mechanism involving specific binding sites in the area postrema of the brain, with outflow through the efferent pathways of the vagus nerve to the gastrointestinal system, rather than by direct action on the stomach. Elevated glucagon concentrations favor increased rates of hepatic glucose release, and a reduction in postprandial glucagon concentrations should result in lower rates of hepatic glucose output during the postprandial period, thus favoring lower postprandial glucose concentrations. It has been demonstrated that hypoglycemia overrides these effects.

Immunoassay of amylin in healthy human volunteers indicates fasting concentrations between 4 and 8 pmol/L, increasing two- to three-fold following ingestion of a mixed meal or an oral glucose load. In patients with type 1 diabetes mellitus, amylin concentrations are near or below the limit of quantitation under fasting conditions and do not increase in response to nutrient stimuli. In patients with type 2 diabetes mellitus or with impaired glucose tolerance, fasting amylin concentrations are comparable to those seen in healthy human subjects. However, the postprandial responses vary considerably. The postprandial responses tend to be decreased in relation to the prevailing level of glycemia and are virtually absent in patients with type 2 diabetes who have progressed to insulin therapy. Amylin deficiency in patients with diabetes mellitus may contribute to impaired glucoregulation.

TERMS and ABBREVIATIONS

AA ---------------------------------

Agency ---------------------------

AUC -------------------------------

BA ---------------------------------

BE ---------------------------------

BMI ---------------------------------

C_max -------------------------------

DMEDP --------------------------

OCPB -----------------------------

NDA -------------------------------

T_max --------------------------------

t_1/2 ----------------------------------

Amino acid(s)

Food and Drug Administration

Area under the plasma-concentration-time curve

Bioavailability

Bioequivalence

Body Mass Index

Maximum drug concentration

Division of Metabolic and Endocrine Drug Products

Office of Clinical Pharmacology and Biopharmaceutics

New Drug Application

Time of maximum drug concentration (C_max)

Drug elimination half-life

DRUG CHARACTERISTICS

Drug Chemistry

Pramlintide acetate is a synthetic analogue of the endogenous human polypeptide, amylin. Pramlintide differs from amylin in its replacement of amino-acid (AA) residues at 25 (alanine), 28 (serine), and 29 (serine) of the 37-AA amylin peptide, with proline residues. These substitutions are purported to increase pramlintide’s solubility, and decrease the propensity for aggregation and adhesion, which has been reported with amylin. Pramlintide acetate is an odorless white powder, is soluble in water, has a molecular weight of 3949.9, and a molecular formula of C₁₇₁H₂₆₇N₅₁O₅₃S₂*xC₂H₄O₂, where x is variable. The structural formula is shown below, and includes the disulfide bridge between the two cysteine residues:

Lys-Cys-Asn-Thr-Ala-Thr-Cys-Ala-Thr-Gln-Arg-Leu-Ala-Asn-Phe-Leu-Val-His-Ser-Ser-

Asn-Asn-Phe-Gly-Pro-Ile-Leu-Pro-Pro-Thr-Asn-Val-Gly-Ser-Asn-Thr-Tyr-NH₂ acetate

Drug Formulation

Multiple formulations (i.e., vial = 26; cartridge = 3) have been described in the development of Symlin™. Of these formulations, AC-0137-F22 (vial) and AC-0137-F28 (cartridge) were chosen for marketing. The formulations for both dosage units are presented in TABLE 1:

TABLE 1: Symlin™ Formulations

	Vial – F22	Cartridge – F28
Strength	0.6 mcg/mL (5 mL)	1.0 mcg/mL (1.5 mL)
pH	4.0	4.0
Pramlintide Acetate Mannitol Metacresol Acetic Acid Sodium Acetate Trihydrate Water for Injection	0.60 g/L 30 mM 43 g/L 2.25 g/L 1.53 g/L 0.61 g/L QS to 1.0 L or 1.015 kg	1.00 g/L 30 mM 43 g/L 2.25 g/L 1.53 g/L 0.61 g/L QS to 1.0 L or 1.015 kg*
*Meets both the European Pharmacopeia (EP) and United States Pharmacopeia (USP) monographs

Early pharmacokinetic studies utilized formulations other than those shown above that differed mainly in their relative pH values – those studies were not reviewed. The data that provided the basis for the Clinical Pharmacology and Biopharmaceutics recommendation was generated using the above to-be-marketed formulations.

Additionally, the sponsor is using three suppliers of pramlintide in the manufacture of Symlin™, Bachem, Mallinckrodt, and UCB-Bioproducts (see HUMAN PK – BIOEQUIVALENCE).

ANALYTICAL

Is the analytical method used to detect pramlintide in human plasma precise and accurate?

Is there any assay interference from endogenous substances or metabolites?

The detection of pramlintide is dependent upon a validated immunoenzymetric assay (IEMA) method that relies on two monoclonal mouse antibodies (Ab): a capture Ab (F024-4.4) and a detection Ab (F025-27.4). Neither Ab is specific for amylin or pramlintide. Antibody F024-4.4 binds near the amino terminus of pramlintide and Ab F025-27.4 binds to the amidated carboxy terminus of pramlintide. The F025-27.4 Ab is conjugated to alkaline phosphatase and a fluorescent substrate, 4-methylumbelliferyl phosphate (4-MUP). This method is able to detect bound, conjugated antibody, using a microplate fluorometer. Relative fluorescence units (RFU) are correlated with concentration using a calibration curve defined in the same assay. This assay uses the Tripro-amylin EIA kit produced by PerSeptive Diagnostics, Inc, and a microplate fluorometer from Dynatech, Chantilly, VA.

In TABLE 2, the quality control samples used in the assays for studies 137-142 and 137-145 are presented. These two studies were chosen because they were the two pivotal bioequivalence studies submitted with this application. Both accuracy and precision were within accepted parameters for these studies. However, there is a concern related to the quality control samples.

Quality control samples were measured at the sponsor defined values of low, middle, and high. However, because the extreme values were far away from the limits of the calibration curve, there exists some doubt about the values obtained at the extremes, especially those plasma pramlintide values that fall between the LLOQ and the lowest quality control sample value. Therefore, because of this issue and the cross-reactivity with endogenous amylin, pharmacokinetic analysis will be confined to C_max and AUC_0-t.

Assay interference was documented for endogenous amylin and for des-lysine pramlintide with 35% and 100% cross-reactivity, respectively. This cross-reactivity results in a total pramlintide concentration that is actually a combination of amylin, des-lysine pramlintide, and pramlintide. Since under fasting conditions endogenous amylin concentrations are usually undetectable, and the fact that the des-lysine pramlintide metabolite is equipotent with pramlintide, these findings are not likely to undermine the value of the assay.

TABLE 2 – Assay Quality Control Results from Two Pivotal Bioequivalence Studies

Study		137-142	137-145
LLOQ (pmol/L)		4.5	4.5
Calibration Range		5.4 – 111.1	5.4 – 111.1
Precision (%CV) 12.7 40.6 40.1 74.9 79.5		3.7 2.9 4.1	4.7 4.1 5.0
Accuracy (%) 12.7 40.6 40.1 74.9 79.5		104.4 102.4 105.2	104.2 101.2 101.1
Stability	5 freeze/thaw cycles
Storage	1 year @ either –20°C or –70°C
Crossreactivity	35% – amylin; 100% – des-lysine pramlintide
Homology	92% – amylin; 97% – des-lysine pramlintide

One important point. This IEMA utilizes monoclonal antibodies from a murine source – and, many humans carry the human anti-mouse antibody (HAMA). HAMA, present in plasma, binds to the mouse antibodies in the assay, which can result in a false signal, either positive or negative. Strategies for reducing this interference include: immunosuppressant therapy, and the use of humanized, polyethylene glycolated, or Fab fragments of antibody agents. The sponsor chose to reduce the HAMA interference by using a diluent that contained a high concentration of non-specific mouse IgG which competes with the specific mouse monoclonal antibodies for the human anti-mouse antibody binding sites.

The above point was raised because several subjects were determined to be unevaluble in several PK studies due to this assay interference. The Agency strongly encourages the sponsor to develop an assay that does not utilize murine antibodies.

HUMAN PK – BIOAVAILABILITY/BIOEQUIVALENCE

Symlin is a parenterally (i.e., subcutaneously) administered sterile solution. As such, knowing the absolute bioavailability is crucial. It is also imperative that there is a good understanding of the effect that physiochemical alterations have on the product’s bioavailability. Therefore, there are two primary questions that should be asked:

What is the absolute bioavailability of Symlin™?

What effect does pH, mixing, volume, or concentration have on the bioavailability of Symlin?

Absolute Bioavailability

The absolute bioavailability for Symlin was determined in an open-label, randomized, four-period crossover study (137-125) in 40 (39 completed[i]) healthy male subjects between 18 and 41 years of age. The four treatments consisted of: A) 60 mcg (0.1 mL) pH 4.7 pramlintide administered subcutaneously (formulation F11); B) 60 mcg (0.1 mL) pH 4.0 pramlintide administered subcutaneously (formulation F22); C) 60 mcg (0.2 mL) pH 4.0 pramlintide administered subcutaneously (formulation F21); and D) 60 mcg (0.1 mL) pH 4.0 pramlintide administered intravenously (formulation F22). Each of the treatments was separated by a 1-week washout period.

Results, presented in TABLE 3 and PLOT 1, show that the absolute bioavailability (f) of a 60 mcg dose of pramlintide, in a volume of 0.1 mL, administered to healthy volunteers under fasting conditions, is 37%. However, because the sampling times for the IV administered pramlintide did not include a sufficient number of early time points, the absolute bioavailability of pramlintide may be overestimated.

TABLE 3 – Pramlintide Plasma Concentrations: SC vs. IV Administration

Parameter

Units

Tx B: 60 mcg (0.1 mL) SC pH 4.0

Tx D: 60 mcg (0.1 mL) IV pH 4.0

Mean

%CV

Mean

%CV

C_max

T_max

AUC_0-t

AUC_0-inf

t_1/2

pmol/L

minutes

pmol*min/L

minutes

89.23

20.5

6234

6803

43.6

24.01

6.51

2154

2253

13.6

26.9

31.9

34.6

33.1

31.2

17,950

18,420

33.4

6834

6795

7.94

38.1

36.9

23.8

C_max / BMI

AUC_0-t / BMI

AUC_0-inf / BMI

pmol/L

pmol*min/L

3.90

272.0

297.2

1.19

99.39

103.6

30.5

36.5

34.9

789.4

809.6

325.4

324.4

41.2

40.1

Analysis was also conducted to determine if body mass index (BMI) had any impact on the coefficient of variation for both C_max and AUC. Results do not indicate a significant difference between the uncorrected parameters and those corrected for BMI. This suggests that BMI has little influence on the pharmacokinetics of pramlintide in healthy males.

Text Box:

Effect of pH on Bioavailability

Formulation pH was shown to have a significant effect on the bioavailability of subcutaneously administered pramlintide. In study 137-125, a pH 4.7 formulation (F11) was compared with the to-be-marketed pH 4.0 formulation (F22). Results, as shown in TABLE 4, revealed that the pH 4.7 formulation had a 25% reduction in AUC and a 35% decrease in the mean C_max, when compared with the pH 4.0 formulation. Because of this pH issue, earlier studies which utilized the pH 4.7 formulation were not considered in this review.

TABLE 4 – Pramlintide Plasma Concentrations: pH 4.7 vs. pH 4.0

Parameter

Units

Tx A: 60 mcg (0.1 mL) SC pH 4.7

Tx B: 60 mcg (0.1 mL) SC pH 4.0

Mean

%CV

Mean

%CV

C_max

T_max

AUC_0-t

AUC_0-inf

t_1/2

pmol/L

minutes

pmol*min/L

minutes

59.70

20.7

4485

5193

49.3

24.01

13.2

2206

2208

18.9

40.2

63.8

49.2

42.5

38.3

89.23

20.5

6234

6803

43.6

24.01

6.51

2154

2253

13.6

26.9

31.9

34.6

33.1

31.2

C_max / BMI

AUC_0-t / BMI

AUC_0-inf / BMI

pmol/L

pmol*min/L

2.62

195.6

226.4

1.15

100.3

99.89

43.9

51.3

44.1

3.90

272.0

297.2

1.19

99.39

103.6

30.5

36.5

34.9

Effect of Volume on Bioavailability

The effect of the volume of SC injections on pramlintide PK was described in several studies, with the most indicative examples being the single/multiple dose studies conducted in type 1 and type 2 diabetes patients. In study 137-143, patients with type 1 DM were administered either a 0.03 mL, 0.06 mL, or 0.09 mL of the 1.0 mg/mL cartridge formulation. In study 137-144, patients with type 2 DM were administered either a 0.06 mL, 0.09 mL, 0.12 mL, or 0.18 mL of the 1.0 mg/mL cartridge formulation. In both cases, AUC_0-300
min and C_max values exhibited expected dose related changes (see HUMAN PK – TARGET POPULATION). Similar results were observed in the absolute BA study, 137-125, where a 0.1 mL dose of the 0.6 mg/mL vial formulation was compared with 0.2 mL of a 0.3 mg/mL “test” formulation. These combined results suggest that the volume of the injection has no significant impact on the bioavailability of Symlin.

Effect of Mixing on Bioavailability

The idea of mixing is an important consideration with Symlin, as it will be used as an adjunct to insulin therapy. With this in mind, the sponsor has conducted numerous studies in which Symlin was mixed in the same syringe with different insulin products (e.g., regular insulin, insulin lispro, NPH insulin, etc.). The overwhelming conclusion from reviewing these studies is that Symlin should not be mixed in the same syringe with insulin because Symlin PK is significantly affected and efficacy will likely be compromised (see HUMAN PK – EXTRINSIC – Drug-Drug Interactions).

Also, due to the importance of pH on the bioavailability of Symlin, mixing Symlin in the same syringe with any agent that can potentially alter the pH is not recommended.

Effect of Concentration on Bioavailability

The effect of concentration on bioavailability was evaluated in two studies, 137-125 and 137-142. Study 137-125 compared a single 60 mcg dose at a concentration of 0.1 mL at pH 4.0 (F22 – to be marketed formulation – reference) with a single 60 mcg dose at a concentration of 0.2 mL at pH 4.0 (F21 – test). Results indicate that a doubling of the injection volume, and hence a halving of the concentration, results in a 6.5% reduction in the relative BA (AUC_0-t).

Study 137-142, discussed in HUMAN PK – Bioequivalence, demonstrated that a 1.0 mg/mL formulation (cartridge) was bioequivalent to a 0.6 mg/mL formulation (vial). It also showed that the relative bioavailability of the cartridge formulation to the vial formulation was about 97% (AUC_{0-300 min}).

Therefore, small changes in drug concentration do not appear to have a significant impact on the bioavailability of Symlin.

Dose Proportionality

Does Symlin™ exhibit dose proportionality over the entire proposed dosing range, 30 mcg to 180 mcg?

In order to characterize dose proportionality, the sponsor has conducted a randomized four-way crossover study in 40 (38 completed) healthy male subjects using four single subcutaneous doses of pramlintide administered at constant volume (0.2 mL). The treatments were as follows: Tx A – 30 mcg (as 0.1 mL of formulation F21 (lot 96-0201GB) plus 0.1 mL placebo (lot 95-0504GE); Tx B – 60 mcg (as 0.1 mL of formulation F22 (lot 96-0503JB) plus 0.1 mL placebo (lot 95-05804GE); Tx C – 90 mcg (as 0.1 mL of formulation F24 (lot 96-0506JB) plus 0.1 mL placebo (lot 95-0504GE); and Tx D – 120 mcg (as 0.2 mL of formulation F22(lot 96-0503JB). Each treatment phase was separated by a 1-week washout period.

Results (TABLE 5) indicate that SC administered pramlintide exhibits near linear kinetics between 30 and 120 mcg in normal healthy subjects – with a definite dose related increase in the PK parameters (see Plots 2 & 3). However, this study failed to demonstrate dose proportionality.

TABLE 5 – Pramlintide PK Profiles in Normal Healthy Subjects – Single Dose

Parameter

Units

30 mcg

60 mcg

90 mcg

120 mcg

C_max

AUC_0-300

T_max

t_½

Cl (apparent)

pmol/L

pmol*min/L

min

L/min

39.26 ± 9.21

3215 ± 1122

21.4 ± 8.79

54.9 ± 14.5

2.36

79.44 ± 20.50

6261 ± 2401

19.5 ± 7.69

49.2 ± 15.3

2.42

102.48 ± 30.23

7939 ± 2848

19.1 ± 7.70

51.1 ± 20.0

2.87

146.99 ± 35.50

11380 ± 3839

21.3 ± 8.36

48.1 ± 12.8

2.67

Mean ± SD

Text Box:

Two additional studies, 137-143 and 137-144, considered dose linearity and dose proportionality in type 1 and type 2 diabetes patients, respectively. Results were similar to those seen in study 137-126, with linearity being demonstrated between the range of 30 and 180 mcg (see HUMAN PK – TARGET POPULATION).

HUMAN PK – BIOEQUIVALENCE

Given that Symlin™ will be available in two concentrations, are these formulations bioequivalent?

Cartridge vs. Vial

Two dosage forms have been proposed for marketing by the sponsor, a cartridge form, to be used with a pen system, and a vial dosage form, to be used for syringe administration. Study 137-142 was conducted to establish bioequivalence between these two dosage forms. In this single center, open-label, randomized, two-period crossover study consisting of two evaluation periods with a 24-hour washout period between dosing events, a single 60 mcg dose administered by pen from the cartridge form (1.0 mg/mL) was compared with a single 60 mcg dose administered by syringe from the vial form (0.6 mg/mL) in 30 subjects (20 females and 10 males). All doses were administered into the subcutaneous tissue of the anterior abdominal wall.

TABLE 6 – Pramlintide PK Parameters: Cartridge Formulation vs. Vial Formulation

Parameter

Units

Tx A: Pen System

60 mcg @ 1.0 mg/mL

Tx B: Vial (syringe)

60 mcg @ 0.6 mg/mL

Mean

%CV

Mean

%CV

C_max

T_max

AUC_{0-300 min}

AUC_0-inf

t_1/2

pmol/L

minutes

pmol*min/L

minutes

71.2

21.8

5742.0

6532.8

52.8

23.75

9.51

3366.3

3499.7

16.51

33.4

43.6

58.6

53.6

31.3

65.7

25.2

5695.5

6704.2

59.4

21.89

15.17

3489.35

3725.05

26.55

33.3

60.2

61.3

55.6

44.7

TABLE 7 – Pramlintide BE Comparison: Cartridge Formulation vs. Vial Formulation

Parameter		Units	Tx A: Pen System	Tx B: Vial (syringe)	Tx: A/B Ratio^c	90% CI^d		p-value^e
Parameter		Units	Mean^b	Mean^b	Tx: A/B Ratio^c	Low	High	p-value^e
C_max^a T_max AUC_{0-300 min}^a AUC_0-inf^a t_1/2		pmol/L minutes pmolmin/L pmolmin/L minutes	67.3 20.0 4845 5673 52.8	62.1 20.0 4679 5673 59.4	108.4 103.6 100.0	100.0 93.3 90.5	117.4 115.0 110.5	0.0986 0.1226^f 0.5742 0.9993 0.1554
a – b – - - c – d – e – f –	Parameters were natural log-transformed before analysis. Means for the test (pen) and reference (syringe) treatment formulations; (geometric means – anti-log of the means of the logs – for natural log-transformed parameters); and (arithmetic means are presented for t_1/2 and median values presented for T_max). Ratio of geometric means calculated as Test/Reference. 90% CI of the geometric means ratio T/R. P-value from ANOVA (sequence, subject-within-sequence, period, & treatment) for testing treatment differences. P-value from the Wilcoxon signed-rank test for the difference of 2 treatment formulations.

Results of this study indicate that the cartridge and vial formulations are bioequivalent when 60 mcg is administered subcutaneously to healthy individuals. This study also suggests that small differences in concentration, i.e., 1.0 mg/mL (cartridge) vs. 0.6 mg/mL (vial), results in non-significant differences in the rate of absorption and no detectable difference in the extent of absorption for this product (see PLOT 4).

Text Box:

Since there will be three suppliers of pramlintide material for Symlin™, are there any PK-related concerns about using multiple sources of this protein?

UCB-Bioproducts vs. Bachem Material

There has been a great deal of concern about the adequacy of the chemical characterization of peptides and their associated process impurities. Since the sponsor is proposing that three independent suppliers of pramlintide provide material for Symlin, the Agency has requested comparative bioavailability information. Specifically, a comparison was requested between USB-Bioproducts material, which has never been used in clinical studies, and either Mallinckrodt or Bachem material, both of which have been used extensively in clinical development and are thought to be equivalent.

As such, the sponsor has submitted a two-way crossover design study (137-145) conducted in 30 normal healthy male and female subjects. Treatments consisted of a single 60 mcg dose from a 1.0 mg/mL formulation in cartridge form with active ingredient manufactured by UCB-Bioproducts [AC137-F28 (99-0603KB) that was compared with a single 60 mcg dose from a 1.0 mg/mL formulation in cartridge form with active ingredient manufactured by Bachem [AC137-F28 (99-0602KB)].

Results of this study (see TABLE 8) clearly show that the F28 cartridge formulation produced from the UCB-Bioproducts pramlintide material is bioequivalent to the F28 cartridge formulation produced from the Bachem material.

TABLE 8 – Pramlintide BE Comparison: UCB-Bioproducts vs. Bachem Pramlintide Material

Parameter		Units	Tx A: UCB	Tx B: Bachem	Tx: A/B Ratio^c	90% CI^d		p-value^e
Parameter		Units	Mean^b	Mean^b	Tx: A/B Ratio^c	Low	High	p-value^e
C_max^a T_max AUC_{0-300 min}^a AUC_0-inf^a t_1/2		pmol/L minutes pmolmin/L pmolmin/L minutes	50.4 17.8 2463.1 3631.6 55.8	51.8 16.2 2554.4 3609.9 56.9	0.97 0.96 1.01	89.4 85.6 89.6	105.9 108.6 113.1	0.5817 0.4747 0.6067 0.9279 0.8657
a – b – - - c – d – e –	Parameters were natural log-transformed before analysis. Means for the test (UCB-Boproducts) and reference (Bachem) treatment formulations; (geometric means – anti-log of the means of the logs – for natural log-transformed parameters); and (arithmetic means are presented for t_1/2 and T_max). Ratio of geometric means calculated as Test/Reference. 90% CI of the geometric means ratio T/R. P-value from ANOVA (sequence, subject-within-sequence, period, & treatment) for testing treatment differences.

HUMAN PK – TARGET POPULATION

Are there any differences between the pramlintide PK profiles of type 1 and type 2 diabetes patients?

Single and multiple dose PK/PD studies were conducted in both type 1 and type 2 diabetes patients. Common findings to both of these studies were: dose linearity over the study-specific dosage ranges, similar confounding factors that led to inconclusive pharmacodynamic conclusions (e.g., incomplete insulin usage records), and no apparent dose accumulation between treatments days 1 and 5. Of note, is the observation that concentrations in type 2 patients tend to be lower than those seen in type 1 patients.

Study 137-143 assessed the single and multiple dose PK profiles of SC administered pramlintide in 11 type 1 diabetes patients. This study was a randomized, three-treatment, three-way crossover design with two dosing frequency groups: three times daily (TID) and four times daily (QID). The group 1 treatments consisted of either 30 mcg, 60 mcg, or 90 mcg pramlintide doses administered SC using the 1.0 mg/mL cartridge formulation and given just prior to breakfast, lunch, and dinner for 4 days followed by a single dose prior to breakfast on the 5^th day (13 consecutive doses). The group 2 treatments were similar to the group 1 treatments, but with the addition of a fourth dose administered just prior to an evening snack (17 consecutive doses). Results are presented in TABLE 9.

TABLE 9 – Pramlintide PK Profiles in Type 1 Diabetes Patients – Single & Multiple Dosing

Parameter	Day	30 mcg TID	60 mcg TID	90 mcg TID	30 mcg QID	60 mcg QID	90 mcg QID
C_max (pmol/L)	1	41.9 ± 22.9	64.5 ± 23.7	99.4 ± 31.4	36.5 ± 10.2	66.9 ± 25.0	123.9 ± 40.5
C_max (pmol/L)	5	37.6 ± 22.8	74.4 ± 20.0	92.7 ± 26.7	40.7 ± 20.0	70.7 ± 23.3	98.6 ± 30.1
T_max (hr)	1	0.273 ± 0.0753	0.365 ± 0.237	0.321 ± 0.115	0.274 ± 0.0751	0.276 ± 0.0794	0.273 ± 0.072
T_max (hr)	5	0.328 ± 0.120	0.334 ± 0.113	0.288 ± 0.083	0.249 ± 0.002	0.278 ± 0.080	0.275 ± 0.082
AUC_0-t *(pmolhr/L)**	1	32.86 ± 28.40	74.14 ± 34.26	118.5 ± 65.37	20.10 ± 11.46	51.67 ± 33.37	115.9 ± 54.29
AUC_0-t *(pmolhr/L)**	5	26.45 ± 21.93	79.68 ± 48.98	104.8 ± 63.53	24.20 ± 18.76	63.88 ± 49.83	102.0 ± 41.77
AUC_0-inf *(pmolhr/L)**	1	81.37 ± 22.17	102.6 ± 37.4	144.2 ± 65.74	48.88 ± ?	72.01 ± 37.82	129.1 ± 55.74
AUC_0-inf *(pmolhr/L)**	5	64.96 ± 9.63	117.2 ± 44.39	136.3 ± 66.70	66.43 ± 15.99	98.34 ± 56.22	124.1 ± 46.12
t_1/2 (hr)	1	1.12 ± 1.01	0.970 ± 0.337	0.774 ± 0.281	0.595 ± ?	0.656 ± 0.169	0.645 ± 0.165
t_1/2 (hr)	5	0.726 ± 0.239	0.789 ± 0.326	0.722 ± 0.230	0.713 ± 0.173	0.724 ± 0.303	1.02 ± 1.02
k_el (1/hr)	1	0.887 ± 0.374	0.796 ± 0.278	1.03 ± 0.423	1.16 ± ?	1.13 ± 0.316	1.14 ± 0.284
k_el (1/hr)	5	1.02 ± 0.282	1.00 ± 0.384	1.03 ± 0.281	1.00 ± 0.243	1.07 ± 0.342	1.03 ± 0.496
Mean ± SD

Study 137-144 assessed the single and multiple dose PK profiles of SC administered pramlintide in 12 type 2 diabetes patients. This study was a randomized, three-treatment, three-way crossover design with two dosing frequency groups: two times daily (BID) and three times daily (TID). The group 1 treatments consisted of either 60 mcg, 120 mcg, or 180 mcg pramlintide doses administered SC using the 1.0 mg/mL cartridge formulation (lot # 97-0403KB) and given just prior to breakfast and dinner for 4 days followed by a single dose prior to breakfast on the 5^th day (9 consecutive doses). The group 2 treatments included 60 mcg, 90 mcg, and 120 mcg doses of pramlintide administered just prior to breakfast, lunch, and dinner (13 consecutive doses). Results are presented in TABLE 10.

TABLE 10 – Pramlintide PK Profiles in Type 2 Diabetes Patients – Single & Multiple Dosing

Parameter	Day	60 mcg BID	120 mcg BID	180 mcg BID	60 mcg TID	90 mcg TID	120 mcg TID
C_max (pmol/L)	1	50.8 ± 20.3	117.4 ± 116.4	151.3 ± 67.2	36.4 ± 18.5	55.7 ± 25.7	74.0 ± 24.6
C_max (pmol/L)	5	55.9 ± 24.6	97.8 ± 34.0	137.3 ± 36.4	42.2 ± 20.2	60.1 ± 24.4	77.2 ± 28.2
T_max (hr)	1	0.331 ± 0.107	0.341 ± 0.105	0.368 ± 0.221	0.290 ± 0.072	0.253 ± 0.008	0.260 ± 0.030
T_max (hr)	5	0.284 ± 0.072	0.324 ± 0.124	0.339 ± 0.224	0.255 ± 0.009	0.301 ± 0.105	0.275 ± 0.071
AUC_0-t *(pmolhr/L)**	1	62.39 ± 61.98	143.0 ± 139.9	179.4 ± 129.9	32.24 ± 46.23	53.93 ± 52.15	76.24 ± 54.37
AUC_0-t *(pmolhr/L)**	5	61.18 ± 52.32	123.1 ± 97.40	189.0 ± 147.8	35.06 ± 38.05	59.66 ± 46.52	91.19 ± 77.00
AUC_0-inf *(pmolhr/L)**	1	119.1 ± 86.42	201.6 ± 160.0	201.0 ± 140.4	201.5 ± 111.6	124.9 ± 85.32	119.2 ± 79.45
AUC_0-inf *(pmolhr/L)**	5	133.7 ± 67.68	163.4 ± 111.0	232.1 ± 166.2	107.7 ± 67.12	129.2 ± 63.29	152.4 ± 102.7
t_1/2 (hr)	1	1.24 ± 0.774	1.24 ± 0.832	0.926 ± 0.496	2.72 ± 1.49	1.38 ± 1.20	1.10 ± 0.778
t_1/2 (hr)	5	1.52 ± 1.09	1.02 ± 0.502	0.981 ± 0.406	1.42 ± 1.40	1.43 ± 0.991	1.15 ± 0.543
k_el (1/hr)	1	0.699 ± 0.301	0.760 ± 0.408	0.907 ± 0.385	0.300 ± 0.165	0.725 ± 0.338	0.811 ± 0.338
k_el (1/hr)	5	0.594 ± 0.249	0.828 ± 0.369	0.814 ± 0.297	0.791 ± 0.425	0.615 ± 0.243	0.686 ± 0.231
Mean ± SD

Comparison of tables 6 and 7 would suggest that, on average, patients with type 2 diabetes tend to exhibit lower total exposure than do patients with type 1 diabetes, when administered the same dose. It should also be noted that there is incredible variation in the PK parameters. This variability may be due to the route of administration, administration technique, and/or differences in body type.

Pramlintide is renally eliminated, what effect does renal insufficiency have on pramlintide PK?

Effect of Renal Insufficiency on Pramlintide PK

The sponsor conducted a study in order to evaluate the effect of renal insufficiency on pramlintide PK, based upon a rat nephrectomy model, a study that showed that basal levels of amylin were significantly higher in lean, non-diabetic patients with renal failure on chronic hemodialysis, and the fact that diabetes is a leading cause of renal disease.

Study 137-127 evaluated the PK of pramlintide in an open-label, parallel design study in 21 type 1 DM patients with varying degrees of renal function. Individuals were stratified by renal function into 1 of 4 categories based on creatinine clearance (CrCl): CrCl ³ 90 mL/min; CrCl 60-89 mL/min; CrCl 30-59 mL/min; and CrCl < 30 mL/min. Each of the patients self-administered a single 60 mcg dose of pramlintide into the subcutaneous tissue of the anterior abdominal wall.

Based on the results of this study (TABLE 11), the sponsor concluded that “impaired renal function has no significant influence on pramlintide PK; therefore no dosing adjustment is required.” However, it should be noted that because of the high inter-subject variability and the few number of patients enrolled in this parallel design study, there can be no definitive conclusion regarding dosing adjustment (i.e., this study is inconclusive). Therefore, dosing of patients with compromised renal function should be individualized based on efficacy and tolerability.

TABLE 11 – Renal Insufficiency – Plasma Pramlintide Parameters in Type 1 DM Patients

AUC_0-inf (pmol*min/L)

Group I

Group II

Group III

Group IV

Mean ± SD

Min, max

8396 ± 4360

2080, 12346

5128 ± 2655

1764, 9199

10538 ± 7064

3100, 19877

5650 ± 4197

850, 8627

C_max (pmol/L)

84.17 ± 30.69

52.4, 138.3

60.70 ± 30.58

33.2, 112.5

86.10 ± 57.02

31.9, 139.8

60.93 ± 35.00

26.1, 96.1

Mean ± SD

Min, max

T_max (min)

34.2 ± 23.06

10, 75

21.9 ± 8.43

10, 30

30.0 ± 20.41

15, 60

15.0 ± 5.00

10, 20

Mean ± SD

Min, max

t_½ (min)

49.4 ± 10.85

32, 63

64.3 ± 54.78

27, 198

80.7 ± 21.94

55, 105

49.4 ± 35.11

19, 88

Mean ± SD

Min, max

Cl/F

169.7 ± 147.9

75, 443

236.1 ± 139.78

100, 523

135.5 ± 111.65

46, 297

438.2 ± 559.71

107, 1084

Mean ± SD

Min, max

Group I = CrCl ³ 80 mL/min; Group II = CrCl 50-80 mL/min; Group III = CrCl 30-50 mL/min; Group IV = < 30 mL/min

Drug-Drug Interactions

Symlin and Insulins

Can Symlin™ be mixed in the same syringe with insulin(s)?

The primary indication for Symlin™ is as an adjunct to insulin therapy. Since both pramlintide and insulin are administered via the subcutaneous route, for convenience (e.g., fewer daily injections), it would be ideal if the two could be mixed in the same syringe. With this in mind, the sponsor has conducted 4 studies (137-130, 137-115, 137-119, & 137-120) in which type 1 diabetes patients were administered pramlintide and insulin(s) at the same time but in different syringes, or pramlintide and insulin(s) in the same syringe. The insulins that were evaluated in these studies included: Regular, NPH, Lente, Ultralente, & 70/30 (see APPENDIX).

The results of these studies, as a whole, strongly indicate that when pramlintide is administered in the same syringe with any insulin product, that both pramlintide and insulin pharmacokinetics can be substantially altered (see TABLES 12 & 13). This would then preclude the practice of “mixing” for reasons of compromised pramlintide and insulin efficacy. This conclusion is in line with that of the sponsor. The labeling will clearly indicate that Symlin™ should not be mixed with any insulin product (see LABELING COMMENTS).

TABLE 12 – Plasma Pramlintide PK: Symlin™ Mixed with R or NPH Insulin vs. Alone

Parameters	Units	A	B	C
		30 mcg Symlin + R Insulin	30 mcg Symlin + NPH Insulin	30 mcg Symlin

C_max AUC_0-t T_max t_1/2	pmol/L pmol*min/L min min	37.30 ± 20.38 1669 ± 1383 22.8 ± 9.83 51.7 ± 14.6	33.21 ± 20.20 1513 ± 1200 32.3 ± 15.4 45.4 ± 13.8	42.43 ± 20.32 1626 ± 1160 18.5 ± 9.07 47.1 ± 28.6
Mean ± SD

TABLE 13 – Comparison: Symlin mixed with R or NPH Insulin vs. Alone

Parameters	Units	A vs. C		B vs. C
		PE	90% CI	PE	90% CI
C_max AUC_0-t	pmol/L pmol*min/L	86.2 98.9	73.1 – 99.4 76.7 – 121.1	77.2 91.1	64.0 – 90.3 68.9 – 113.3
Mean ± SD

Symlin and Lo/Ovral or Ampicillin

Since pramlintide delays gastric emptying time, what effect does Symlin™ have on orally administered medications?

As previously described, one of pramlintide’s primary mechanisms of action is to delay gastric emptying time. Therefore, it is conceivable that if an orally administered drug were administered concomitantly with pramlintide, an interaction may occur such that the object drug’s efficacy could be compromised. With this in mind, the sponsor has conducted two drug-drug interaction studies to determine the effect that pramlintide has on an oral contraceptive (OC), Lo/Ovral™ and a relatively acid-stable antibiotic, ampicillin.

Study 137-133 evaluated the effect of pramlintide on the PK of ethinyl estradiol and norgestrel in healthy female subjects receiving Lo/Ovral (30 mcg ethinyl estradiol + 300 mcg norgestrel). In this randomized, two-period crossover design study, 18 females on OC treatment were subcutaneously administered either a placebo injection (lot # 96-0302JE) or a 90 mcg dose of Symlin™ (lot # 96-0503JB – F22 – 0.6 mg/mL) 15 minutes before administration of a single dose of Lo/Ovral™ (Lot # 9978046, expiration: 3/2000).

Results as presented in TABLE 14 showed no statistically significant differences in the PK profile of ethinyl estradiol on any of the calculated PK parameters. In contrast, the norgestrel component of Lo/Ovral™ did exhibit significant differences when administered with Symlin™ (see TABLE 15). The C_max for norgestrel was reduced by about 30% and the time to C_max was delayed by 45 minutes; AUC was similar between treatments.

TABLE 14 – Effect of Symlin on Lo/Ovral (ethinyl estradiol) Pharmacokinetics

ETHINYL ESTRADIOL	Lo/Ovral	Lo/Ovral + Symlin	p-value^a	Ratio or LSMeans Difference	95% CI
					Low	High
lnAUC_{0-24 hours} (pg*min/L)	10.4 (0.1) 32990.2	10.4 (0.1) 33076.8	0.981	100.3%	79.4%	126.6%
LS Means (SE)^a GEO LS Means^b
*lnAUC_0-inf (pgmin/L)**	10.8 (0.1) 49544.0	10.9 (0.1) 54227.8	0.389	109.5%	88.1%	136.0%
LS Means (SE)^a GEO LS Means^b
lnC_max (pg/L)	4.6 (0.1) 101.2	4.6 (0.1) 99.7	0.842	98.5%	84.3%	115.2%
LS Means (SE)^a GEO LS Means^b
T_max (min)	74.0 (14.1)	101.9 (14.1)	0.176	27.9	-13.9	69.7
LS Means (SE)^a
t_½ (min)	413.1 (64.4)	468.8 (64.4)	0.412	55.6	-84.8	196.0
LS Means (SE)^a
^a = Based on ANOVA model which includes terms for sequence, subject within sequence, period and treatment ^b = Geometric means are the antilogs of the means of the natural logarithmic transformed endpoints.

TABLE 15– Effect of Symlin on Lo/Ovral (norgestrel) Pharmacokinetics

NORGESTREL	Lo/Ovral	Lo/Ovral + Symlin	p-value^a	Ratio or LSMeans Difference	95% CI
					Low	High
lnAUC_{0-24 hours} (ng*min/L)	7.4 (0.1) 1658.3	7.4 (0.1) 1628.6	0.646	98.2	90.4%	106.7%
LS Means (SE)^a GEO LS Means^b
*lnAUC_0-inf (ngmin/L)**	8.0 (0.1) 3070.2	8.1 (0.1) 3288.8	0.188	107.1	96.3%	119.2%
LS Means (SE)^a GEO LS Means^b
lnC_max (ng/L)	1.5 (0.1) 4.7	1.5 (0.1) 3.2	< 0.001	68.9	60.4%	78.5%
LS Means (SE)^a GEO LS Means^b
T_max (min)	68.9 (11.7)	114.2 (11.7)	0.014	45.3	10.9	79.6
LS Means (SE)^a
t_½ (min)	1780 (205)	2054 (205)	0.197	274.9	-162	712
LS Means (SE)^a
^a = Based on ANOVA model which includes terms for sequence, subject within sequence, period and treatment ^b = Geometric means are the antilogs of the means of the natural logarithmic transformed endpoints.

The second drug-interaction study (137-134), which evaluated delayed gastric emptying, was a double-blind, placebo-controlled, single-dose, two-period crossover design study in 12 (11 completed) healthy male and female subjects. Subjects were given either placebo (lot # 96-032JE) plus a single 2 x 250 mg oral dose of ampicillin (batch # 54763 B; expiration 8/2002) or a single 90 mcg subcutaneously administered dose of Symlin™ (lot # 96-0503JB – F22 – 0.6 mg/mL) plus oral ampicillin. Results (TABLE 16) show that pramlintide has no significant impact on the AUC or C_max of ampicillin, but similar to the Lo/Ovral™ study, T_max was increased by approximately one hour.

TABLE 16 – Effect of Symlin on Ampicillin Pharmacokinetics

AMPICILLIN	Ampicillin	Ampicillin + Symlin	p-value^a	Ratio or LSMeans Difference	95% CI
					Low	High
lnAUC_{0-8 hours} (mcg*min/L)	6.7 (0.1) 782.3	6.6 (0.1) 748.9	0.553	95.7	81.5%	112.4%
LS Means (SE)^a GEO LS Means^b
*lnAUC_0-inf (mcgmin/L)**	6.7 (0.1) 792.9	6.6 (0.1) 1.6 (0.1)	0.559	95.8	81.5%	112.5%
LS Means (SE)^a GEO LS Means^b
lnC_max (mcg/L)	1.5 (0.1) 4.7	1.6 (0.1) 5.1	0.380	108.3	89.0%	131.9%
LS Means (SE)^a GEO LS Means^b
T_max (min)	110.0 (12.7)	176.2 (12.7)	0.003	66.2	29.1	103.3
LS Means (SE)^a
t_½ (min)	59.1 (3.2)	51.0 (3.2)	0.046	-8.2	-16.2	-0.2
LS Means (SE)^a
^a = Based on ANOVA model which includes terms for sequence, subject within sequence, period and treatment ^b = Geometric means are the antilogs of the means of the natural logarithmic transformed endpoints.

Individually, the results from these studies are relatively unremarkable – especially given that the extent of absorption was not affected by pramlintide in either case. However, both of these agents, ampicillin and Lo/Ovral™, are used chronically and are not dependent upon rapid drug action, as are many oral pain medications, etc. Therefore, as the sponsor has recommended, concomitant medications susceptible to delayed gastric emptying times should be administered at least one-hour before administration of Symlin™.

HUMAN PK – PLASMA/BLOOD

An in vitro protein binding study of pramlintide in animal and human blood and plasma. A mean of 37% of pramlintide tracer spiked into whole blood was bound to cells, with the remaining in the unbound in the plasma fraction. Thirty-three percent was bound to soluble plasma components, which leaves approximately 40% total pramlintide available for receptor binding.

HUMAN PK – EX VIVO – Placental Transfer

Placental transfer was evaluated in placentas from normal term vaginal or cesarean section deliveries. Placentas were perfused as an open, non-circulating, system for 90 to 120 minutes and then as a closed, circulating, system for an additional 90 to 120 minutes. Concentrations in the perfusate ranged from 206 to 458 pmol/L.

No detectable pramlintide was detected on the fetal side after 120 minutes of the open-loop perfusion stage. One sample of fetal perfusate, during the closed loop phase, had detectable pramlintide at 60 minutes, but was undetectable at 90 minutes. These ex vivo results suggest that fetal exposure to pramlintide in utero is low.

PHARMACODYNAMICS

What effect does a morning dose of pramlintide have on the gastric emptying of a lunchtime meal?

Effect of Pramlintide on Gastric Emptying

Delay gastric emptying time is one of the two well described mechanisms of pramlintide action. Gastric emptying, as a pharmacodynamic (PD) endpoint, was formally evaluated in two studies (also see HUMAN PK – EXTRINSIC – Drug-Drug Interactions). The first of these studies, 137-118, determined the dose-response relationship of single doses of pramlintide on the rate of gastric emptying of the liquid and solid components of a radio-labeled standardized meal (500 kcal – 55% CHO; 35% fat, 10% protein) and to determine if a dose of pramlintide administered 15 minutes before breakfast has a “carry-over” effect on the emptying of the lunch time meal. This four-way cross-over study in 14 (11 evaluable) type 1 diabetes patients compared treatments of 30 mcg, 60 mcg, and 90 mcg (lot # 95-0902GB – F21 – 0.3 mg/mL formulation) with placebo (lot # 95-0504GE).

Results of this study showed that when pramlintide is administered 15 minutes before a standardized breakfast consisting of a liquid component, 3-ortho-methyl-glucose (3-OMG) labeled milkshake, and a solid meal, ^99mTechnetium-amberlite resin labeled pancake, gastric emptying time was significantly delayed compared with placebo. The time to maximum plasma 3-OMG concentration (T_max) was increased from about 80 minutes for placebo to 200 minutes for the 90 mcg dose (see TABLE 17). The time to achieve half-gastric emptying was similarly increased (see TABLE 18). There was no observed “carry-over” effect of the morning pramlintide dose to the lunch time meal.

TABLE 17 – Liquid Meal – Plasma 3-OMG PK Parameters by Treatment (n = 11)

Strength

Symlin and Insulins

Symlin and Lo/Ovral or Ampicillin