11

95% confidence interval for the risk ratio is greater than 1.8, the guidance states that there

12

would be inadequate evidence of cardiovascular safety to support marketing. If the upper

13

bound falls less than 1.8 and there are no other approvability issues, the drug would be

14

able to be approved but the drug may need a post-marketing required cardiovascular trial

15

to definitively address cardiovascular safety. The cut point for this decision is 1.3. If the

16

upper bound of the risk ratio is between 1.3 and 1.8 post-marketing cardiovascular trials

17

would be needed to show definitively that the upper bound is less than 1.3. If the upper

18

bound for this risk ratio is less than 1.3 and the drug can otherwise be approved, post-

19

marketing cardiovascular trials generally will not be necessary. These upper bounds of

20

1.3 and 1.8 assume a reassuring point estimate.

4

the upper bound represents the “worst case” potential for increased cardiovascular risk

5

based on a combined analysis across studies, and we set some criterion as to what that

6

upper bound should be. The numbers, we will talk about in a moment, are 1.3 and 1.8.

7

For example, an upper bound of 1.8 says that the estimated increased risk of

8

cardiovascular events with an investigational drug is no worse than 1.8 times the risk

9

with comparative.

22

of the applicants. Initially Bristol-Myers Squibb and Novo Nordisk were asked to

23

conduct cardiovascular analyses and they use different approaches and different

24

methodologies to analyze cardiovascular events. To create a more uniform approach

10

Let’s take a closer look at these two numbers, if the upper bound of this

21

I would now like to turn to the cardiovascular analyses that FDA requested

Scribes, LLC Toll Free 1-800-675-8846 www.scribesllc.com

1

compute the point estimate for the risk ratio comparing investigational drug to

2

comparator and also the 95% confidence interval for this risk ratio.

3

19

We then turn to the upper bound of this 95% confidence interval because