DATE: February 14, 2007
TO: Edward Wolfgang, HFM-475
Martha Lee, Ph.D.
A. Dale Horne, Dr. P.H.
Vaccines Evaluation Branch
Pentacel (HCPDT-IPV+PRP-T vaccine)
HFM-475/Edward Wolfgang (original)
HFM-217/A. Dale Horne
In the original BLA 125145 submission, Clinical Summary (Item 8) Section of Pivotal Studies, “Serology Bridging Study for the Pertussis Response in the Pentacel - Clinical Trial 494-01 in the United States and the Sweden Efficacy Trials,” results of a matching analysis are presented in the Final Serology Bridging Study Report (page 50-51 of 321). The matching analysis procedure was performed by the following steps: 1) subjects in Sweden I (N=80) were sorted by pre-immunization PRN value; 2) a subset (n=240) of the 494-01 Stage II data was identified to match in a 3:1 ratio with the subjects in Sweden I by the same pre-immunization PRN value; 3) the seroconversion rate for PRN was calculated from this subset of subjects; 4) the subset of 240 subjects was returned to the whole pool of 494-01 subjects (N=444); 5) steps 2,3, and 4 were repeated 9,999 times for a total of 10,000 samples; and 6) the seroconversion rates from the 10,000 samples were averaged, and that averaged seroconversion rate was used for comparing to the Sweden I seroconversion rate. This matching analysis had been proposed in IND ---------- and reviewed by Dr. Henry Hsu in December, 2004. A copy of Dr. Hsu’s review is attached to the end of this memo.
Summary of the Statistical Review by Dr. Henry Hsu
The sponsor performed a post hoc “matching” analysis of the seroconversion rates from the serology bridging study in order to take into account an uneven distribution of pre-immunization antibody levels between the Sweden CPDT and the 494-01 Pentacel groups. Results of the matching analysis showed that the adjusted seroconversion rate for the Pentacel group based on 240 subjects was 93.2%, an increase of 4% from 89.2%. The difference [CPDT – Pentacel] in rates was 5.60% (please note that it is 5.55% in the BLA submission) with an upper limit of the 95% CI of 9.56%, just meeting the non-inferiority criterion of the upper limit of the 95% CI being < 10%.
Comments and concerns regarding the methods and results include:
- It is a post-hoc analysis after the results of the pre-specified analysis failed the non-inferiority criterion set in advance.
- The result is an indirect adjusted seroconversion rate based on a pre-selected four-fold rise criterion in the simulation analysis. Hence, this result does not provide adequate evidence to support the non-inferiority claim.
- It is not clear why the n=3 matching number (i.e., in the ratio 3:1) was chosen. It might be more informative to have several different matching numbers, because different matching numbers may provide different scenarios. It would also be reasonable to have simulation results that are based on the Sweden CPDT subset with PRN pre-immunization antibody levels similar to the Pentacel group.
- Result of the adjusted ratio of GMTs is not presented.
In both 494-01 and P3T06 studies, results regarding the pertussis PRN antigen are consistently lower in the Pentacel group than in the separate injection control group. It is not clear whether the levels of PRN pre-immunization would be higher in the Pentacel group than in the control group and what the differences would be if a similar matching adjustment were made.
Recommendation: The sponsor proposed the same matching analysis in the BLA submission without new information. Our comments and concerns remain in effect.