The following is excerpted from the June 2005 deposition testimony of Deborah Shapiro, senior biostatistician for Merck & Co. In the testimony, regarding a clinical analysis that linked Vioxx with vascular adverse effects, Shapiro acknowledges that certain information was not given to the U.S. Food and Drug Administration (FDA). New Jersey Superior Court Judge Carol Higbee called the testimony the only evidence sufficient to support a punitive damages verdict against Merck. Questioning Shapiro for the plaintiffs is Shanin Specter of Kline & Specter, PC.

20 Now, let me ask you about what you
21 gave to the FDA in terms of the meta-analysis. And
22 that was Exhibit 82, and I will put that in front of
23 you.
24 Now, this document has 61 numbered
25 pages; is that correct? Actually, it has two pages

1 at the start that constitute the cover letter to the
2 FDA from Merck and then a form as the second page,
3 and then the document entitled "Interim
4 Cardiovascular Meta-Analysis" has 61 numbered pages.
5 A. Yes.
6 Q. Is that all fair?
7 A. Yes.
8 Q. Now, I want to talk about how this is
9 organized for a minute, if we could.
10 Section Roman numeral IV which is on
11 Page 12 is titled "Results." Is that correct?
12 A. Yes.
13 Q. The first thing that appears there is
14 A, which is "Summary." Correct?
15 A. Yes.
16 Q. And then three pages later at Page 15
17 we have Section B of results, which are "APTC
18 Endpoint." Correct?
19 A. Yes.
20 Q. Then the remainder of the Results
21 section discusses the APTC endpoint; correct?
22 A. Even the introductory part does. It
23 is just a summary of all those analyses.
24 Q. Was there a part C?
25 A. There's parts A and B.

972
1 Q. I know. I see there are parts A and
2 B. I'm asking a different question now. Was there
3 a part C?
4 A. No.
5 Q. Was there ever a part C?
6 A. I don't recall.
7 Q. Let's see if I can refresh your
8 recollection.
9 MR. SPECTER: Mark this as
10 Shapiro-88. A copy for you, a copy for your lawyer.
11 (Handing over documents.)
12 - - -
13 (Whereupon, Deposition Exhibit
14 Shapiro-88, "Full Meta-Analysis,"
15 MRK-ACF0000845 - MRK-ACF0000924, was
16 marked for identification.)
17 - - -
18 (Witness reviewing document.)
19 BY MR. SPECTER:
20 Q. Please tell us what you're looking
21 at.
22 A. I just need to look through it to
23 figure it out.
24 Q. Go ahead.
25 A. (Witness reviewing document.)

973
1 Q. Doctor, what are you looking at right
2 now?
3 A. Can you let me finish looking,
4 please?
5 Q. I want to know what you are looking
6 at. What are you looking at?
7 A. I'm looking at a section C. I would
8 like to continue looking, please.
9 Q. Fine. What page is it on?
10 A. Bates Number ending 909.
11 Q. Thank you.
12 A. (Witness reviewing document.)
13 Q. Let's compare these two documents.
14 Let's start off first with the question I left off
15 with, which is, is there a section C? Is there a
16 section C?
17 A. There's a section C. It is not
18 reflected on the table of contents, I believe. No,
19 it's not in the table of contents, so, there's a
20 section labeled C with a blank page in between it.
21 Q. Right. And this entire exhibit
22 that's been marked as Shapiro-88 is consecutively
23 Bates numbered; correct?
24 A. Seems to be.
25 Q. Let's compare the two documents,

974
1 Exhibit 82, which is the one that your lawyer showed
2 you that was sent to the FDA, and Exhibit 88, which
3 is the one that I put in front of you.
4 If we start off on numbered Page 1 of
5 the document your lawyer gave you, that's titled
6 "Interim Cardiovascular Meta-Analysis correct?
7 A. Yes.
8 Q. And if we look at numbered Page 1 on
9 what I just gave you, it says the same thing,
10 "Interim Cardiovascular Meta-Analysis."
11 A. Yes.
12 Q. Is that correct?
13 A. Yes.
14 Q. And then the documents are identical
15 for the first 61 pages; correct?
16 A. I haven't reviewed all 61 pages. If
17 you say they are identical, I'm hoping they are, but
18 it would take me a long time to compare them.
19 Q. Well, you sort of went through them
20 somewhat quickly just now; is that correct?
21 A. Yes, I did, very quickly.
22 Q. Did you see any differences in the
23 first 61 pages?
24 MR. MAYER: You can answer the
25 question whether you saw any differences.

975
1 THE WITNESS: I didn't see any
2 differences, but I didn't read them thoroughly, so,
3 it is hard for me to say there's no differences.
4 BY MR. SPECTER:
5 Q. That's fine. However, on the
6 document I just showed you, there is a page before
7 the statement that appears "Interim Cardiovascular
8 Meta-Analysis," and that page is labeled "Full
9 Meta-Analysis." Correct?
10 A. No. Where? What page number?
11 Q. Just before the page labeled "Interim
12 Cardiovascular Meta-Analysis," there's a page and
13 all that's on it is "Full Meta-Analysis." Right
14 here. (Indicating.)
15 A. Can you tell me what page number you
16 are on, Bates Number, please.
17 Q. It is literally -- Bates Number
18 MRK-ACF0000845.
19 A. Thank you.
20 Q. It is literally the first page of the
21 exhibit; right?
22 A. Yes.
23 Q. What does it say?
24 A. "Full Meta-analysis."
25 Q. "Full Meta-analysis."

976
1 Now, the thing given to the FDA, has
2 that got a page before the page that says "Interim
3 Cardiovascular Meta-Analysis"? Does that have a
4 page before that that says "Full Meta-Analysis"?
5 A. No.
6 Q. But the one I just gave you has that;
7 correct?
8 A. Yes.
9 Q. It has that page that says "Full
10 Meta-Analysis." Right?
11 A. Yes.
12 Q. Then it's got a section C, and the
13 title of section C is "MI Endpoint - Secondary."
14 Correct?
15 A. Yes.
16 Q. And then there is discussion and then
17 a whole bunch of tables analyzing the data from the
18 standpoint of the MI endpoint; correct?
19 A. Yes.
20 Q. There is no section C in what was
21 given to the FDA; correct?
22 A. Yes.
23 Q. In fact, the bottom line of the page,
24 the first page of the discussion of the MI endpoint,
25 the last sentence is, "Whereas there was a

977
1 borderline significant difference in treatment
2 effects between comparisons with naproxen and other
3 NSAIDs for the APTC endpoint, this difference was
4 significant for the MI endpoint."
5 A. I believe that's --
6 MR. MAYER: Object to the form.
7 THE WITNESS: -- referring to
8 heterogeneity, which is more apparent for MI. So,
9 that's actually stronger evidence of not combining.
10 But you will have to give me a couple of minutes to
11 look at that to make sure that's what it is
12 referring to. That's my initial belief, that it's
13 the heterogeneity significance that that's referring
14 to.
15 BY MR. SPECTER:
16 Q. What question are you answering now,
17 Dr. Shapiro?
18 A. Yours.
19 Q. What question was that? What was the
20 question?
21 A. That one had a borderline and one had
22 a significant.
23 Q. I just read to you what it said;
24 correct?
25 A. Okay.

978
1 Q. Is that correct?
2 A. I was going beyond and interpreting.
3 And I'm sorry.
4 Q. Did I read what appears there
5 correctly?
6 A. I would have to hear you read it
7 again, but I think so.
8 Q. I'll read it again for you. "Whereas
9 there was a borderline significant difference in
10 treatment effects between comparisons with naproxen
11 and other NSAIDs for the APTC endpoint, this
12 difference was significant for the MI endpoint."
13 Did I read it correctly?
14 A. Yes.
15 Q. Thank you.
16 Then we have a whole bunch of tables
17 on this subject; correct?
18 A. Yes.
19 Q. Do these tables appear in what was
20 given to the FDA?
21 A. These tables do not appear.
22 Q. Did that sentence I just read appear
23 in what was given to the FDA?
24 A. No.
25 Q. For example, looking at Appendix 1.3,

979
1 which appears at Bates Number 923 where it talks
2 about "Meta-Analysis of MI Endpoint Vioxx versus
3 Comparator Agents," and it goes to the VIGOR trial,
4 and it has 20 heart attacks in the Vioxx group and 4
5 on the comparator with a relative risk of 5.0. Did
6 I read that correctly?
7 A. Yes.
8 Q. And then in combining one, two,
9 three, four, five, six studies of people with
10 rheumatoid arthritis with the meta-analysis of the
11 MI endpoint, the relative risk is 4.16. Is that
12 what it says?
13 A. Yes. That's the analysis where the
14 heterogeneity means you can't combine it, and I
15 believe it was footnoted in one of these tables,
16 that you can't combine it because that was any NSAID
17 versus placebo.
18 Q. I'm looking, yeah, at that, Doctor.
19 Where is that footnote on that page that says you
20 can't combine it? Show me where that footnote
21 appears there.
22 A. It doesn't appear there, but if you
23 actually repeat the sentence that you said to me,
24 that's talking about heterogeneity, and
25 heterogeneity implies the inability to combine.

980
1 Q. Doctor, the sentence that I read to
2 you, where does the word "heterogeneity" appear in
3 that sentence?
4 A. You have to understand that a
5 comparison between naproxen and nonnaproxen reflects
6 an interaction in heterogeneity. So, it's something
7 that the FDA would understand. It doesn't say
8 heterogeneity, but that's the implication.
9 Q. So, this Appendix 1.3 "Meta-Analysis
10 of MI Endpoint Vioxx versus Comparator Agents," was
11 this table given to the FDA with the meta-analysis?
12 A. No, it's not provided --
13 Q. Right.
14 A. -- but MI information was provided
15 for each and every comparison. The MI information
16 is in each table that breaks out results for each
17 comparison.
18 Q. Oh, you mean that Table 4 that you
19 talked about in answering your lawyer's questions;
20 right?
21 MR. MAYER: Object to the form.
22 THE WITNESS: Tables 4, 6, 8, 10 and
23 12.
24 BY MR. SPECTER:
25 Q. Let's go to Table 4.

981
1 A. (Witness complies.)
2 Q. So, Table 4, "Cardiac Events," "Acute
3 Myocardial Infarction," that's a heart attack;
4 right?
5 A. Yes.
6 Q. 28 people taking Vioxx, 6 people
7 taking naproxen; right?
8 A. Yes.
9 Q. Where is the statement there about
10 relative risk? Is there a statement of relative
11 risk there?
12 A. There's not a statement of relative
13 risk, but there are and --
14 Q. Doctor, don't "but" me. Just answer
15 my question.
16 MR. MAYER: Objection.
17 BY MR. SPECTER:
18 Q. Is there a statement of relative
19 risk?
20 MR. MAYER: Let her finish her
21 answer.
22 BY MR. SPECTER:
23 Q. Is there -- please answer the
24 question, Dr. Shapiro.
25 Is there a statement of relative risk

982
1 in the table?
2 A. While there's not a statement of
3 relative risk, one can take that plus the row that
4 says myocardial infarction and derive a relative
5 risk estimate from the percentages. The data is not
6 hidden. It is very clearly there.
7 Q. Dr. Shapiro, in order to determine a
8 valid relative risk, you would have to have the
9 information as to patient years; correct? Correct?
10 A. You -- well, you don't -- yes, in a
11 form, but you still can --
12 Q. Doctor, don't "but" me.
13 You have to have the patient years to
14 have relative risk; correct?
15 MR. MAYER: I object to the continued
16 interruptions of the witness' answers. Let her
17 answer the question and ask the question you want to
18 ask her.
19 THE WITNESS: There's more than one
20 way to derive a relative risk. One uses patient
21 years, that assumption there that risk is constant
22 over time. There are several other ways to derive
23 relative risk. You can derive relative risk from a
24 COX model. You can derive relative risk from a
25 ratio of rates, and that is often done. One can

983
1 take the percentages and divide them, and that is
2 also considered a relative risk.
3 BY MR. SPECTER:
4 Q. And neither was done for the FDA in
5 what was sent to them? You didn't give them the
6 relative risks either with or without patient years;
7 correct? Yes or no?
8 A. We provided the raw ingredients that
9 anyone can derive a relative risk from. It's quite
10 simple.
11 Q. Anyone can do it?
12 A. Yes.
13 Q. Do you think I can?
14 A. I would hope so. You did some pretty
15 good math before.
16 Q. Do you think the jurors can in this
17 case?
18 A. I think they can get a pretty good
19 idea.
20 Q. You think the jurors can take 0.31
21 and 0.08 and determine relative risk from that?
22 MR. MAYER: Object to the form.
23 THE WITNESS: I think they can
24 certainly get an estimate of one being higher and by
25 how much from that. I'm sure they are capable of

984
1 looking at those numbers and seeing a difference.
2 BY MR. SPECTER:
3 Q. Okay. Well, we'll see if they agree
4 with that.
5 Looking at Appendix 1.3, though,
6 Doctor, in what you didn't give the FDA, you had the
7 relative risks there, didn't you?
8 MR. MAYER: Object to the form.
9 THE WITNESS: We had the relative
10 risks.
11 BY MR. SPECTER:
12 Q. And you had the patient years there,
13 don't you?
14 A. Yes.
15 Q. Doctor, when I handed you Exhibit 88
16 a few minutes ago, and you saw there was a section
17 C, were you surprised to see that? Yes or no?
18 A. I didn't remember it.
19 Q. Please answer my question.
20 When I showed it to you and showed
21 you that there was a section C with an MI endpoint,
22 were you surprised to see it? Yes or no?
23 A. Well, I know I had looked at MIs and
24 had analyzed them. So, hard to say whether I was
25 surprised. I didn't recall writing it up, but I

985
1 recall doing the analyses. So...
2 Q. Doctor, whether or not it would be
3 hard to say whether you were surprised, were you, in
4 fact, surprised to see this document with a section
5 C?
6 A. No. I -- the APTC endpoint
7 encompasses it. The results in terms of
8 heterogeneity were stronger for the MI endpoint.
9 It's one piece of the APTC. We weren't hiding the
10 MI results. We provided the APTC endpoint because
11 you can get more precision with more events, the
12 events track together. We pointed out to the FDA
13 that what was driving it were the MI results. It's
14 all there, and it's all listed. It's --
15 Q. It's all there, and it's all listed?
16 A. Yes.
17 Q. I see. The relative risks are listed
18 and there for the FDA in the meta-analysis; is that
19 correct?
20 A. The FDA can very easily divide two
21 numbers.
22 Q. And these tables that appear at the
23 back of the section on results of the MI endpoint in
24 what you at Merck label a, quote, Full
25 meta-analysis, that's there for the FDA in what you

986
1 gave them as the meta-analysis? Yes or no?
2 MR. MAYER: Objection to the form.
3 THE WITNESS: Section C is not
4 included --
5 BY MR. SPECTER:
6 Q. Right.
7 A. -- but all the information to
8 evaluate drug effects and the different comparators
9 is there for them.
10 Q. Dr. Shapiro, did you ever see Exhibit
11 88 before I just handed it to you?
12 A. Well, considering I wrote it, yes.