Understanding the 505(b)(2) Approval Pathway

Download Report

Transcript Understanding the 505(b)(2) Approval Pathway

Charles Bon
19 May 2015
Two-Way, Randomized Crossover Study
12-80+ Healthy, Normal Adults
-48 to -12 Hour Check-in
Overnight Diet and Activity Restrictions
Single Dose (1/2 subj. get Test & 1/2 get Ref)
10-Hour Pre-Dose to 4-Hour Post-Dose Fast
15-25 Samples collected over 3-4 half-lives
Adequate washout (crossover design)
Crossover to Alternate Product (usual)

Drug Concentration Time Profile
Measured Drug Concentrations
AUCt = Area Under the Curve 0-t
Sum ( ½ * (C1 + C2) * (T2 - T1) )
Calculated to Ct
AUCinf = AUCt + Ct / Ke
Proportional to amount absorbed
Can calculate only if we have Ke
Measured Drug Concentrations
AUCt , AUC - Extent of Absorption
Cmax - Rate & Extent of Absorption
Tmax - Rate of Absorption
Terminal Rate of Elimination (Ke , ß, z )
Terminal Half-Life of Elimination (t½)
Test of Equality (not of use for BE)
Standard ANOVA (=0.05)
H0: New = Standard
Ha: New  Standard
H0: New / Standard = 1
Ha: New / Standard  1
BE Requires a Test of Comparability
Two, One-Sided T-tests (=0.05)
H01: New / Standard < LL
Ha1: New / Standard  LL
H02: New / Standard > UL
Ha2: New / Standard  UL
Average BE
- ≤ (T - R) ≤ 
 = Ln(1.25)
- = -Ln(1.25) = Ln(0.80)
2, 1-sided t-test ( = 0.05)  2-sided 90% CI
Same As: (T - R)2 ≤ 2
One-Sided 95% UCB
• Highly Variable Drugs
• Narrow Therapeutic Index Drugs
• In-Vitro Population BE
Guidance for Industry
Statistical Approaches to Establishing
Bioequivalence
U.S. Department of Health and Human Services Food and Drug
Administration Center for Drug Evaluation and Research (CDER)
January 2001
BP
APPENDIX A
E(T - R)2
E(R1 -
R2)2
 
Analysis of Ln-Transformed AUC & Cmax
Mixed-Effects Linear Model
Each subject, j, provides µTj and µRj
µTj & µRj from Distn(µT) & Distn(µR)
σBT2 and σBR2
Correlation, ρ, between µTj and µRj.
σD2 is related to these parameters
σD2 = variance of (µTj - µRj)
= (σBT - σBR)2 + 2 (1-ρ)σBTσBR
The total variances for each formulation are defined as the sum of the
within- and between-subject components
σ
TT
2=σ
2+σ
2
WT
BT
2+σ 2
σ 2=σ
RR
WR
BR
For analysis of crossover studies, the means are given additional
structure by the inclusion of period and sequence effect terms.
A mixed-scaling approach was suggested for individual
BE. Reference-scaled method if the estimate of σWR >
σW0, constant-scaling otherwise, with σW0 = 0.20 (σWR2
or σW02 as denominator). The guidance recommends that
I = 0.05
The guidance recommends that sponsors applying the
individual BE approach may use either referencescaling or constant-scaling at either side of the
changeover point.
First Problem
(T - R)2 > Ln(1.25)
Offset by (σBT2 - σBR2)
Must Constrain (T - R)
Second Problem
A subject-by-formulation interaction could occur
when an individual is representative of subjects
present in the general population in low
numbers, for whom the relative BA of the two
products is markedly different than for the
majority of the population, and for whom the
two products are not bioequivalent, even though
they might be bioequivalent in the majority of
the population.
Must constrain σD2
Irreconcilable Problems
Industry Resisted 4-way Studies
FDA’s Influential Proponent Left FDA
Kill the Concept
The Real Problem
The equation for the statistic could be
readily understood by
non-statisticians
High Variability In a Drug
1.
2.
3.
BCS Class III or IV (Low Solubility)
Formulation Effects (MR vs. IR)
Biological Variable (Oral Progesterone)
What Doesn’t Cause High Variability
Analytics
Var(A + B) = Var(A) + Var(B)
Analytics = (0.18)2 Biological = (0.35)2
Var(A + B) = (0.39)2
Analytics = (0.05)2 Biological = (0.35)2
Var(A + B) = (0.36)2
Highly Variable Drugs
Prior to 2008 (2009)
 Don’t pursue product
 Run huge two-way BE study
 Run slightly less huge four-way
2008+
 Replicate design/large study
 Scaled Average BE (USA)
CV = 35%, T/R = 0.93 (1.075), Prob ≥ 0.80
ABE (2-way)
SABE (3-way)
SABE (4-way)
N = 66 (132 SP sets)
N = 30 (90 SP sets)
N = 20 (80 SP sets)
Must Replicate Reference Product
 Partial Replicate: TRR, RTR, RRT
 Full Replicate: TRTR, RTRT
CVWR determines BE method
 ≥ 30% the SABE method is used
 < 30% must use ABE method
Three-way, Crossover Study: TRR, RTR, and RRT.
Sequence, Period and Treatment fixed effects.
Subjects nested within Sequence random effect.
Multiplicative model.
ABE uses ANOVA with above terms.
SABE uses only Sequence in its ANOVA model for
variables (T – R) and (R1 - R2)
What Happens to the Subject Effect?
Subj Effect
Test
Ref
Test/Ref ln(Test/Ref)
0.01
0.9
1
0.9
-0.1054
0.1
9
10
0.9
-0.1054
1
90
100
0.9
-0.1054
10
900
1000
0.9
-0.1054
100
9000
10000
0.9
-0.1054
Using ln (Test/Ref) = ln(Test) – ln(Ref) removes
Subject Effect
What happens to the Period Effect?
T – 0.5 (R1 + R2)
TRR: P1 - 0.5 (
P2 + P3)
RTR: P2 - 0.5 (P1 +
P3)
RRT: P3 - 0.5 (P1 + P2
)
Sum: P1+P2+P3 – P1 -P2-P3
Mean: 0 for period effect
Estimate of (T – R) is  (T – 0.5 (R1 + R2)) / 3
R1 – R2
Var(R1 – R2) = Var(R) + Var(R) = 2Var(R)
Var(R + c) = Var(R)
TRR:
RTR:
RRT:
MSE :
S2 = Var(R + P2) + Var (R + P3) = 2Var(R)
S2 = Var(R + P1) + Var (R + P3) = 2Var(R)
S2 = Var(R + P1) + Var (R + P2) = 2Var(R)
Estimate of 2Var(R)
Estimate of
2
 WR
= MSE/2
SABE Statistics
ANOVA with only sequence term for:
dlat = T – 0.5*(R1+R2)
ilat
= R1- R2
Swr2 = MSE/2 from ilat ANOVA = estimate of wR2
2
2
se dlat ,
x = dlat unbiased estimate of (µT-µR)
y = - s2  , estimate of 2WR 
2
Bound(x) = 95% UCB on dlat2 = (max [Abs (90% CL on dlat)])2
Bound(y) = 95% UCB on y = y*df /crit(0.95,df)
(µT-µR) - WR  ≤ 0
2
2
95% UCB for (µT-µR)2- 2WR 
critbound =
½
2
2
(x+y)+( (boundx-x) +(boundy-y) )
BE concluded if:
• 95% UCB ≤ 0
• 0.80 ≤ T/R ≤ 1.25
Draft Guidance on Progesterone
Active ingredient:
Form/Route:
Progesterone
Capsule/Oral
SAS Program Statements for Scaled Average BE
Analysis of Replicated Crossover Studies
Draft Guidance on Progesterone
PROC MIXED for fully replicated (4-way)?
PROC GLM for partially replicated (3-way)
SAS Program Statements for Average BE
Analysis of Replicated Crossover Studies
PROC MIXED;
CLASSES SEQ SUBJ PER TRT;
MODEL Y = SEQ PER TRT/
DDFM=SATTERTH; RANDOM
TRT/TYPE=FA0(2) SUB=SUBJ G;
REPEATED/GRP=TRT SUB=SUBJ;
ESTIMATE 'T vs. R' TRT 1 -1/CL ALPHA=0.1;
The Problems
 Study not conducted as a single dosing group
 Analyze each group separately
 Group*Seq as only term in statistical model
 Analyze ilat with Goup*Seq / dlat separately
 ABE – nonconvergence of Proc Mixed
 Use CSH, FA0(1)
 ?
Narrow Therapeutic Index Drugs
Draft Guidance on Warfarin Sodium
Active ingredient: Warfarin Sodium
Form/Route:
Tablet/Oral
Narrow Therapeutic Index Drugs
4-period, 2-sequence replicated crossover study
1. Evaluate by ABE (90% CI)
2. Evaluate by SABE with  = Ln(1.111)2/(0.10)2
3. 95% confidence: wt /wR ≤ 2.5
Narrow Therapeutic Index Drugs
Problems


Swr2 can become very small in a given data set
What is considered a NTI drug?
Draft Guidance on Budesonide
Active ingredient: Budesonide
Form/Route:
Suspension/Inhalation
References
Guidance for Industry, “Statistical Approaches to Establishing Bioequivalence”, U.S.
FDA, CDER, Jan. 2001.
Haidar SH, et. al. Bioequivalence approaches for highly variable drugs and drug
products. Pharm Res 2008; 25:237-241.
Haidar SH, et. al. Evaluation of a Scaling Approach for the Bioequivalence of highly
variable drugs. Pharm AAPS J 2008; 10:450-454.
Draft Guidance on Progesterone. U.S. FDA, CDER, Feb 2011.
Draft Guidance on Warfarin Sodium. U.S. FDA, CDER, Dec 2012.
Draft Guidance on Budesonide. U.S. FDA, CDER, Sep 2012.
Thank You