• Decrease font size
  • Return font size to normal
  • Increase font size
U.S. Department of Health and Human Services

Vaccines, Blood & Biologics

  • Print
  • Share
  • E-mail

Statistical Review and Evaluation - Evicel

DEPARTMENT OF HEALTH AND HUMAN SERVICES
PUBLIC HEALTH SERVICE
FOOD AND DRUG ADMINISTRATION
CENTER FOR BIOLOGICS EVALUATION AND RESEARCH
Division of Biostatistics (HFM-215)

STATISTICAL REVIEW AND EVALUATION

Date Populated:  Jan. 14, 2008
Type/Application ID/Amendment #:  BLA 125010/105
Proposed Use (Indication):  Adjunct to hemostasis
Sponsor:  Omrix.
Product name:  EVICEL fibrin sealant (Human)
From:  Chinying Wang, Ph. D. (HFM-219)
Through:  Ghanshyam Gupta, Ph. D.
To:  Debbie Cordaro (HFM-380)
Cc:  Ghanshyam Gupta, Ph. D., Henry Hsu, Ph. D.,
Chron. File (HFM -215)

The study protocol entitled "A Prospective, Randomized Controlled Study to Compare the Effects of a Fibrin Sealant (FS2) as an adjunct to hemostasis for soft tissue bleeding during retroperitoneal or intra-abdominal surgery" is a phase III study comparing FS2 to Surgicel. This efficacy supplement contains study reports in support of a proposed new indication for adjunct to hemostasis in patients undergoing non-emergent retroperitoneal or intra-abdominal surgery procedures where a topical hemostat is indicated.

Statistical Review

The efficacy analysis for the primary endpoint has been reviewed. Efficacy was defined as FS2 was non-inferior to Surgical. The primary efficacy endpoint is defined as an absence of bleeding and no treatment with any hemostatic measures (other than the assigned hemostatic adjunct, FS2 or Surgicel) at 10 minutes following randomization to treatment. The analysis of primary endpoint is confirmed.

For the analysis of primary endpoint, the sponsor used a two-sided 95% confidence interval (CI) for the ratio of proportions of success. In the statistical analysis plan, it states that in case that the lower limit of the 95% CI is greater than 1.0, then there is evidence of superiority. Statistically, the procedure is agreeable. However, for the labeling of superiority of FS2 to Surgical, there is a concern for the conservativeness due to the original study design of non-inferiority. Using Fisher exact test (shown in the appendix), there is statistical difference between FS2 and Surgicel. The superiority claim would be fine.

In addition, analyses of secondary endpoints for absence of bleeding at 4 and 7 minutes following randomization have been performed. The survival analysis, using Kaplan-Meier curves was performed for time to hemostasis. The survival curves for FS2 and Surgicel are slightly different from that presented by the sponsor because the sponsor considered treatment failure as censored at 10minutes. The analyses are attached in the Appendix.

APPENDIX

(Ia) Contingency Analysis of SUCDC By TREATN (ITT primary endpoint)

TREATN By SUCDC

Count Total % Col % Row %

No

Yes

 

FS2

3
2.22
18.75
4.55

63
46.67 52.94 95.45

66
48.89

Surgicel

13
9.63
81.25 18.84

56
41.48 47.06 81.16

69
51.11

 

16
11.85

119
88.15

135

 

Test

ChiSquare

Prob>ChiSq

Likelihood Ratio

7.084

0.0078

Pearson

6.598

0.0102

Fisher's Exact Test

Prob

Alternative Hypothesis
Left

0.0093

Prob(SUCDC=Yes) is greater for TREATN=FS2 than Surgicel
Right

0.9984

Prob(SUCDC=Yes) is greater for TREATN=Surgicel than FS2
2-Tail

0.0148

Prob(SUCDC=Yes) is different across TREATN

(Ib) Contingency Analysis of SUCDC By TREATN (PP)

TREATN By SUCDC

Count Total % Col % Row %

No

Yes

 

FS2

3
2.36
18.75
4.84

59
46.46
53.15
95.16

62
48.82

Surgicel

13
10.24
81.25
20.00

52
40.94
46.85
80.00

65
51.18

 

16
12.60

111
87.40

127

 

Test

ChiSquare

Prob>ChiSq

Likelihood Ratio

7.109

0.0077

Pearson

6.624

0.0101

Fisher's Exact Test

Prob

Alternative Hypothesis
Left

0.0092

Prob(SUCDC=Yes) is greater for TREATN=FS2 than Surgicel
Right

0.9984

Prob(SUCDC=Yes) is greater for TREATN=Surgicel than FS2
2-Tail

0.0144

Prob(SUCDC=Yes) is different across TREATN

(II) Contingency Analysis of HAE4DC By TREATN

TREATN By HAE4DC

Count
Total %
Col %
Row %

No

Yes

 

FS2

16
11.85
33.33
24.24

50
37.04
57.47
75.76

66
48.89

Surgicel

32
23.70
66.67
46.38

37
27.41
42.53
53.62

69
51.11

 

48
35.56

87
64.44

135

 

Test

ChiSquare

Prob>ChiSq

Likelihood Ratio

7.320

0.0068

Pearson

7.213

0.0072

Fisher's Exact Test

Prob

Alternative Hypothesis
Left

0.0059

Prob(HAE4DC=Yes) is greater for TREATN=FS2 than Surgicel
Right

0.9981

Prob(HAE4DC=Yes) is greater for TREATN=Surgicel than FS2
2-Tail

0.0114

Prob(HAE4DC=Yes) is different across TREATN

(III) Contingency Analysis of HAE7DC By TREATN

TREATN By HAE4DC

Count Total % Col % Row %

No

Yes

 

FS2

6
4.44
27.27
9.09

60
44.44
53.10
90.91

66
48.89

Surgicel

16
11.85
72.73
23.19

53
39.26
46.90
76.81

69
51.11

 

22
16.30

113
83.70

135

 

Test

ChiSquare

Prob>ChiSq

Likelihood Ratio

5.084

0.0242

Pearson

4.915

0.0266

Fisher's Exact Test

Prob

Alternative Hypothesis
Left

0.0226

Prob(HAE7DC=Yes) is greater for TREATN=FS2 than Surgicel
Right

0.9937

Prob(HAE7DC=Yes) is greater for TREATN=Surgicel than FS2
2-Tail

0.0354

Prob(HAE7DC=Yes) is different across TREATN

(IV) Product-Limit Survival Fit

Survival Plot

chart showing survival plot fs2 and surgicel

 

Time to event: HAEMTM         Grouped by TREATN

Summary

Group

N Failed

N Censored

Mean

 

Std Error

FS2

63

0

2.82857

 

0.25678

Surgicel

56

0

3.69821

 

0.26141

Combined

119

0

3.23782

 

0.18689

Quantiles

Group

Median Time

Lower95%

Upper95%

25% Failures

75% Failures

FS2

2.25

1.7167

3.1

1.1333

4

Surgicel

3.3667

2.7167

4

2.25

4.6667

Combined

3

2.4333

3.4333

1.6667

4.2

Tests Between Groups

Test

ChiSquare

DF

Prob>ChiSq

Log-Rank

3.6733

1

0.0553

Wilcoxon

7.3567

1

0.0067