Contains Nonbinding Recommendations
Draft Guidance on Warfarin Sodium
This draft guidance, once finalized, will reprent the Food and Drug Administration's (FDA's) current thinking on this topic. It does not create or confer any rights for or on any person and does not operate to bind FDA or the public. You can u an alternative approach if the approach satisfies the requirements of the applicable statutes and regulations. If you want to discuss an alternative approach, contact the Office of Generic Drugs.
Active ingredient: Warfarin Sodium
Form/Route: Tablet/Oral
Recommended studies: 2 studies
1. Type of study: Fasting
Design: 4-way, fully replicated crossover design in-vivo
Strength: 10 mg
Subjects: Healthy males and nonpregnant females, general population.
Additional Comments: Warfarin has a long terminal elimination half-life. Plea ensure
adequate washout periods between treatments in the crossover studies. For long half-life
drug products, an AUC truncated to 72 hours may be ud in place of AUC0-t or AUC0-∞, as described in the Guidance for Industry: “Bioavailability and Bioequivalence Studies for
Orally Administered Drug Products – General Considerations”.
Applicants may consider using the reference-scaled average bioequivalence approach for
warfarin described below.
2. Type of study: Fed
Design: 4-way, fully replicated crossover design in-vivo
Strength: 10 mg
Subjects: Healthy males and nonpregnant females, general population.
Additional Comments: See additional comments above. See Amantadine Hydrochloride
Oral Tablet guidance for recommendation regarding fed studies.
gym怎么读
Analytes to measure (in appropriate biological fluid): Warfarin in plasma, using an achiral assay. Bioequivalence bad on (90% CI): Warfarin
Waiver request of in-vivo testing: 1 mg, 2 mg, 2.5 mg, 3 mg, 4 mg, 5 mg, 6 mg and
7.5 mg are eligible for a waiver of in-vivo bioequivalence testing bad on (i) acceptable bioequivalence studies on the 10 mg strength, (ii) acceptable in vitro dissolution testing of all strengths, and (iii) proportional similarity of the formulations across all strengths.
Dissolution test method and sampling times:
Plea note that a Dissolution Method Databa is available to the public at the OGD website at www.v/scripts/cder/dissolution/. Plea find the dissolution information for this product at this website. Plea conduct comparative dissolution testing on 12 dosage units each
of all strengths of the test and reference products. Specifications will be determined upon review of the application.
Explanation : FDA has concluded that Warfarin sodium is a narrow therapeutic index (NTI) drug bad on the following evidence:
• For warfarin there is a narrow range between therapeutic and toxic dos or the associated blood
or plasma concentrations (i.e., exposures);
• Warfarin toxicities are rious and not symptomatic or reversible;
• Subtherapeutic warfarin concentrations may lead to rious and life-threatening complications; • Warfarin is subject to therapeutic monitoring bad on pharmacodynamic markers; and
• Warfarin has low within subject variability.
The study should be a fully replicated crossover design in order to
• Scale bioequivalence limits to the variability of the reference product; and
• Compare test and reference product within-subject variability.
Method for Statistical Analysis Using the Reference-Scaled Average Bioequivalence Approach for na
rrow therapeutic index drugs:
Step 1. Determine s WR , the estimate of within-subject standard deviation (SD) of the
reference product, for the pharmacokinetic (PK) parameters AUC and Cmax.
Calculation for can be conducted as follows:greensock
WR s
()()2.1122j
n m ij i i j WR D D s n m ==−=−∑∑
Where:
i = number of quences m ud in the study
[m=2 for fully replicated design: TRTR and RTRT]
j = number of subjects within each quence T = Test product R = Reference product
(where 1 and 2 reprent replicate reference treatments)
1ij ij ij D R R =−2
1.i n ij
j i i D
D n ==∑
(i.e. total number of subjects ud in the study, while is number of
1m
j i n ==∑n i n subjects ud in quence i )
Step 2. U the referenced-scaled procedure to determine BE for individual PK
parameter(s).
Determine the 95% upper confidence bound for:
()22R T W Y Y s θ−−R Where:
中文转英文
• T Y and R Y are the means of the ln-transformed PK endpoint (AUC and/or Cmax)
obtained from the BE study for the test and reference products, respectively
• 20ln()W θσ⎛⎞Δ≡⎜⎟ (scaled average BE limit) ⎝⎠
• and 00.10W σ= (regulatory constant), 1.11111Δ=(=1/0.9, the upper BE limit)
The method of obtaining the upper confidence bound is bad on Howe’s Approximation I, which is described in the following paper:
W.G. Howe (1974), Approximate Confidence Limits on the Mean of X+Y Where X and Y are Two Tabled Independent Random Variables, Journal of the American Statistical Association, 69 (347): 789-794.
Step 3. U the unscaled average bioequivalence procedure to determine BE for
individual PK parameter(s). Every study should pass the scaled average
bioequivalence limits and also regular unscaled bioequivalence limits of 80.00-
125.00%.
Step 4.
Calculate the 90% confidence interval of the ratio of the within subject standard deviation of test product to reference product WT WR σσ. The upper limit of the
90% confidence interval for WT WR σσ will be evaluated to determine if σWT and
σWR are comparable . The propod requirement for the upper limit of the 90%
equal-tails confidence interval for σWT /σWR is less than or equal to 2.5.
The (1)100%α−CI for WT WR
σσ
is given by
where
• WT s is the estimate of WT σwith 1v as the degree of freedom
• WR s is the estimate of WR σ with 2v as the degree of freedom
• 21,,2/F νναis the value of the F-distribution with 1ν (numerator) and
2ν(denominator) degrees of freedom that has probability of 2/αto its
right.
• 21,,2/1 is the value of the F-distribution with 1F ννα−ν (numerator) and
2ν(denominator) degrees of freedom that has probability of 1-2/αto its
right.
• here 0.1α=.
Example SAS Codes: 4-period, 2-quence replicated crossover study
For a bioequivalence study with the following quence assignments in a fully replicated 4-way crossover design:
Period 1 Period 2 Period 3 Period 4
Sequence 1
T R T R Sequence 2
R T R T
The following codes are an example of the determination of reference-scaled average bioequivalenc
e for LAUCT. Assume that the datats TEST and REF, have already been created, with TEST having all of the test obrvations and REF having all of the reference obrvations.
Datat containing TEST 1 obrvations:
data test1;
t test;
if (q=1 and per=1) or (q=2 and per=2);
la run ;
t1t=lauct;
Datat containing TEST 2 obrvations:
data test2;
t test;if (q=1 and per=3) or (q=2 and per=4); l run ;
at2t=lauct;
Datat containing REFERENCE 1 obrvations:
data ref1;
t ref;if (q=1 and per=2) or (q=2 and per=1);
l run ;
at1r=lauct;
Datat containing REFERENCE 2 obrvations:
data ref2;
t ref;
if (q=1 and per=4) or (q=2 and per=3);
lat2r=lauct;
run ;
The number of subjects in each quence is n 1 and n 2 for quences 1 and 2, respectively.
职位分析Define the following quantities:
T k th ijk =obrvation (k = 1 or 2) on T for subject j within quence i R =k th ijk obrvation (k = 1 or 2) on R for subject j within quence i
12122ij ij ij ij ij T T R R I ++=−2
and
明尼苏达大学研究生12ij ij ij D R R =−
I ij is the difference between the mean of a subject’s (specifically subject j within quence i) two obrvations on T and the mean of the subject’s two obrvations on R, while D ij is the difference between a subject’s two obrvations on R.
Determine I ij and D ij
data scavbe;
友谊地久天长英文歌词
merge test1 test2 ref1 ref2;
by q subj;
ilat=0.5*(lat1t+lat2t-lat1r-lat2r);
dlat=lat1r-lat2r;
run;
Intermediate analysis - ilat
proc mixed data=scavbe;
class q;
model ilat =q/ddfm=satterth;
estimate 'average' intercept 1 q 0.5 0.5/e cl alpha=0.1;
ods output CovParms=iout1;
ods output Estimates=iout2;
ods output NObs=iout3;
title1 'scaled average BE';
title2 'intermediate analysis - ilat, mixed';
run;
From the datat IOUT2, calculate the following:
IOUT2:
edfapointest=exp(estimate);
x=estimate**2–stderr**2;
boundx=(max((abs(lower)),(abs(upper))))**2;
Intermediate analysis - dlat
proc mixed data=scavbe;
class q;
model dlat=q/ddfm=satterth;
estimate 'average' intercept 1 q 0.5 0.5/e cl alpha=0.1;
ods output CovParms=dout1;
ods output Estimates=dout2;
ods output NObs=dout3;
title1 'scaled average BE';
t
服装导购员itle2 'intermediate analysis - dlat, mixed';
run;
From the datat DOUT1, calculate the following:
DOUT1: s2wr=estimate/2;年底总结报告
From the datat DOUT2, calculate the following:
DOUT2: dfd=df;陷入困境
From the above parameters, calculate the final 95% upper confidence bound:
theta=((log(1.11111))/0.1)**2;
y=-theta*s2wr;
