Part II: Celexa Pediatric Clinical Trials


Part I | Part II | Part III | Part IV | Part V | Part VI

There were two pediatric trials conducted on Celexa:  Study 94404 and MD-18.  And, as discussed below, both were negative for efficacy on every primary and secondary endpoint.

I. Celexa Study 94404 Was a Negative Clinical Trial

Celexa Study 94404 was conducted in Europe by Forest’s partner, Lundbeck, and was submitted to the FDA on March 21, 2002.[45]  The study involved 244 depressed adolescents, aged 13-18.[46] The study had one primary endpoint and nine secondary endpoints.[47]  As illustrated in Table 1, all endpoints were negative.

Table 1 – Celexa Study 94404 Efficacy Results[48]

Endpoint P-Value Result
Change in Kiddie-SADS-P Total Score over Time (Primary) N/A Negative
Change from Baseline in Kiddie-SADS-P Total Score 0.791 Negative
Kiddie-SADS-P Response 1.000 Negative
Change from Baseline in MADRS Total Score 0.853 Negative
MADRS Response 0.865 Negative
MADRS Remission 0.867 Negative
Beck’s Depression Inventory (BDI) 0.863 Negative
Global Assessment of Functioning (GAF) 0.933 Negative
Life Event Scale N/A Negative
Expressed Emotions N/A Negative

Forest learned that Study 94404 was negative on July 16, 2001.[49] This was approximately around the time Forest also learned about the results of MD-18.  Forest made a deliberate decision to suppress the results of 94404 while promoting the results of MD-18.[50] Dr. William Heydorn, a former Senior Medical Writer at Forest, explained:

  1. Were you aware of anyone at Forest Labs who shared the view that it would be best if the positive data of CIT-MD-18 Was in the marketplace before the negative data of 94404 was out in the marketplace?
  2. Yes.
  1. And who did you understand to share that view?
  2. I think most of the individuals associated with the citalopram project held that view.
  1. Was it your understanding at the time that you were working at Forest Labs that positive data would be better than negative data in terms of marketing Celexa?
  2. Yes.
  1. And that positive data being put out in the marketplace over negative data would be better for the sales of Celexa?
  2. I certainly wasn’t in the sales and marketing department, but that would be my understanding, yes.[51]


The investigators at Lundbeck were eager to get the results of 94404 published, but Forest and Lundbeck wanted to make sure the “positive” results of MD-18 were in the public domain.[52]  Thus, for three years, the results of Study 94404 remained concealed.  Then, in 2004, the New York Times published an article criticizing Forest for publishing MD-18 without mentioning the negative results of Study 94404.[53]  This prompted an immediate response from Forest where it issued a press release disclosing the results of Study 94404.[54]  Study 94404, however, did not get published until 2006—five years after Forest and Lundbeck obtained the results.[55]

II. Celexa Study MD-18 Was a Negative Clinical Trial, but Forest Misled the FDA about the Results

Celexa Study MD-18 was conducted by Forest in the United States.  It involved 174 pediatric patients diagnosed with depression, aged 7-17.[56]  The final study report for MD-18 was submitted to the FDA on April 8, 2002.[57]  MD-18 had one primary and four secondary endpoints.[58]  As illustrated in Table 2, all five endpoints were negative.  For those patients who actually completed MD-18, i.e., “observed cases,” the results were also negative (p-value of 0.167).[59]

Table 2 – Celexa Study MD-18 Efficacy Results [60]

Endpoint P-Value Result
Change from Baseline in CDRS-R at 8 Weeks (LOCF) (Primary) 0.052* Negative
Change from Baseline in CDRS-R at 8 Weeks (Observed Cases) 0.167 Negative
CGI Improvement at 8 Weeks 0.257 Negative
Change from Baseline in CGI Severity at 8 Weeks 0.266 Negative
Change from Baseline in CGAS at 8 Weeks 0.309 Negative
Change from Baseline in K-SADS-P Depression Module at 8 Weeks 0.105 Column 2 Value

*P-value based on data excluding 9 patients who were unblinded during the study.

In any DBRCT, data collected from the patients must be double-blind. The protocol for MD-18 stated that: “Any patient for whom the blind has been broken will immediately be discontinued from the study and no further efficacy evaluations will be performed.”[61] Dr. William Heydorn, the primary author of the final study report for MD-18,[62] confirmed that, pursuant to the protocol, unblinded patients were to be excluded from any efficacy analysis.[63] And, this makes sense. Efficacy in a depression trial is based on an investigator’s subjective assessment of a patient’s subjective responses to questions on a depression questionnaire. If either the patient or investigator knows whether the patient is taking a real drug or a placebo, it could very well influence the patient’s answers and the investigator’s subjective scoring. Indeed, the FDA has identified multiple types of bias associated with blinding:

  • Subjects on active drug might report more favorable outcomes because they expect a benefit or might be more likely to stay in a study if they knew they were on active drug.
  • Observers might be less likely to identify and report treatment responses in a no-treatment group or might be more sensitive to a favorable outcome or adverse event in patients receiving active drug.
  • Knowledge of treatment assignment could affect vigor of attempts to obtain on-study or follow-up data.
  • Knowledge of treatment assignment could affect decisions about whether a subject should remain on treatment or receive concomitant medications or other ancillary therapy.
  • Knowledge of treatment assignment could affect decisions as to whether a given subject’s results should be included in an analysis.
  • Knowledge of treatment assignment could affect choice of statistical analysis.

Blinding is intended to ensure that subjective assessments and decisions are not affected by knowledge of treatment assignment.[64]

Forest told the FDA, DOJ, and USAO that MD-18 was a positive study because, even though all of the secondary endpoints were negative (as well as the observed cases analysis), Forest claimed the primary endpoint reached statistical significance. This, however, was untrue. As discussed in detail below, the first nine patients randomized into MD-18 were unblinded because of a packaging error. Under the protocol, these initial nine patients should not have been included in the efficacy analysis. And, without these patients included, MD-18 was negative on every endpoint, including the primary endpoint. Internal documents and the testimony of former employees demonstrate that Forest deliberately misled the FDA about the extent of the unblinding and that MD-18, when properly assessed, is negative.

A. General Overview of MD-18 Study

Dr. Paul Tiseo, Joan Barton, and Dr. Charles Flicker oversaw MD-18.[65] Dr. Tiseo was the Medical Monitor for MD-18, Ms. Barton was the Clinical Trial Manager of MD-18, and Dr. Flicker was Dr. Tiseo and Ms. Barton’s supervisor, overseeing all of the clinical trial programs related to Celexa and Lexapro.[66]  Dr. Tiseo was responsible for the overall conduct of the study.[67] Above Dr. Flicker was Dr. Ivan Gergel and, above him, was Dr. Lawrence Olanoff.[68]  After MD-18 was completed, Dr. Tiseo left Forest, and Dr. William Heydorn took responsibility for drafting and publishing the results of MD-18 in both the final study report and subsequent academic publication.[69]

When a child was enrolled in the study, the child and their parent were dispensed medication at different pre-specified intervals as reflected below:

Visit Week Dispensing Intervals
Visit 1 Week -1 Patient dispensed 1 bottle containing 10 placebo pills.
Visit 2 Week 0 Patient randomized. Dispensed 1 bottle containing 10 pills.
Visit 3 Week 1 Patient dispensed 1 bottle containing 10 pills.
Visit 4 Week 2 Patient dispensed 2 bottles containing 10 pills.
Visit 5 Week 4 Patient dispensed 1 bottle containing 40 pills.
Visit 6 Week 6 Patient dispensed 1 bottle containing 40 pills.
Visit 7 Week 8 Study completed.[70]

At Visit 1 (week -1), each patient was put through a one-week placebo screening period, also known as a placebo run-in.[71] During this period, the patient was given one week of medication in a 10-pill bottle containing placebo pills.[72] This period was single-blinded—meaning the patient did not know the pills were placebo, only the investigator knew.[73]

At Visit 2 (week 0), patients were assessed to see how they responded to the 1 week placebo-screening period.[74] If they responded, they were not allowed to enter the randomized portion of the trial.[75] The remaining patients were randomized into either the placebo or Celexa group. At this point, each patient’s baseline was established.[76] The randomization was supposed to be double-blind, meaning neither the patient nor the investigator knew which group the patient was assigned to. Each patient was given another 10-pill bottle containing either blinded-placebo or blinded-Celexa.[77]

At Visit 3 (week 1), the investigator conducted another round of assessments to see how, if at all, the patient was responding to treatment.[78] As part of this process, each patient was required to return unused medication and the investigator was required to count the number of remaining pills in the bottle to ensure compliance.[79] At the end of the visit, the patients were dispensed a new 10-pill bottle to last until the next visit, the following week.[80]
At Visit 4 (week 2), like at Visit 3, more assessments were done and the pills were counted. The patients were then dispensed two 10-pill bottles to last the next two weeks.[81]

At Visit 5 (week 4), the patients were given another round of assessments and the pills were counted.[82] At this half-way point in the trial, the investigators were permitted to increase the patient’s dose by double if the patient was not responding.[83] If so, the patient was expected to take two pills instead of one each day.[84] Accordingly, at Visit 5 (week 4), each patient was given a 40-pill bottle, which would last two weeks until the next visit.[85]

At Visit 6 (week 6), more assessments and pill counts were conducted and the patients were dispensed another 40-pill bottle to last two more weeks.[86]

At Visit 7 (week 8) the study was completed and the final assessments were performed. The success of each patient was determined by a comparison of the patient’s improvement (or lack of improvement) between Visit 2 (week 0) and Visit 7 (week 8).[87]

B. At the Beginning of the Trial, a Packaging Error Caused Nine Patients, and their Investigators, to Become Unblinded

Shortly after MD-18 began enrolling patients, Forest learned of a packaging error. According to Dr. Tiseo, two “investigational sites called in to report that some of their patients were receiving white tablets and others were receiving pink tablets.”[88] Forest investigated and,“it was discovered that a number of bottles of ‘active’ medication were mistakenly packed with the pink-colored commercial Celexa® tablets instead of the standard white citalopram tablets used for blinded clinical studies.”[89]

According to an investigation by the clinical supply group within Forest, the 10-pill bottles to be used in the Celexa group did not contain the standard white blinded pills, but contained pink, oval-shaped, Forest-branded, and dose-stamped commercial Celexa® tablets.[90] See photo below.

To correct the packaging error, Dr. Tiseo ensured “all sites were notified of this error by telephone and by fax.”[91] In the fax, Dr. Tiseo informed each investigational site about the packaging error and told each site that the pink pills they were seeing in the patients already randomized were “pink-colored commercial Celexa® tablets instead of the standard white citalopram tablets used for blinded clinical studies.”[92]  Dr. Tiseo explained that “dispensing these tablets would automatically unblind the study.”[93] Dr. Tiseo instructed each investigational site to immediately return the unblinded medication for repackaging.[94]  However, for those patients already randomized, i.e., already receiving the commercial Celexa tablets, Dr. Tiseo instructed each site to keep using the unblinded medication.[95]

The problem, however, is that for those patients already randomized into the study, the patients and the investigators were unblinded. The investigators brought the packaging error to Forest’s attention because some patients were receiving white pills and some were receiving pink ones. When Dr. Tiseo told investigators the pink pills were commercial Celexa tablets, even if the patients somehow did not know the Forest-branded tablets were the active drug, the investigators knew the patients getting the pink pills were getting Celexa and the patients getting white pills were getting inert placebos. Dr. Flicker admitted this during his deposition: “[I]f an investigator were to look . . . at returned medication and he saw that the tablets were pink . . . then I would think the investigator would be able to draw the conclusion that the patient was on active drug.”[96] Dr. Heydorn also conceded: “If an investigator knows which patients are taking branded Celexa and which ones are taking white pills, doesn’t that mean the integrity of the blind was . . . unmistakenly compromised? . . .  It does raise questions about the integrity of the blind, yes.”[97]
Additionally, there is strong evidence that the patients randomized into the Celexa group were also, individually, unblinded.  First, the average improvement of the blinded patients in the study taking Celexa was 21.3 points on the CDRS scale.[98]  However, the average improvement for the unblinded patients given commercial Celexa for the first four weeks was 30.5 points.[99]  This 50% greater improvement in the unblinded Celexa patients, versus the blinded Celexa patients, is strong evidence that those patients or investigators were, in fact, unblinded.

Second, the patients in the Celexa group who were given the commercial Celexa tablets would have been exposed to different color and shaped pills throughout the trial.  Specifically, the pink tablets were only located in the 10 pill bottles, which were only dispensed during the four weeks after randomization. The last four weeks of trial used the 40-pill bottles, which contained the standard, blinded, white pills. However, prior to being randomized, every patient was given white placebo pills during the one week placebo run-in period.  So, this means, these patients would have been given white pills for one week, pink commercial Celexa pills for four weeks, then white placebo-looking pills for four weeks.[100]  This is illustrated in the diagram on the next page.[101]



According to the protocol for MD-18, “[a]ny patient for whom the blind has been broken will immediately be discontinued from the study and no further efficacy evaluations will be performed.”[102]  This means any unblinded patients should have been excluded from the study efficacy evaluations.  Dr. Heydorn confirmed this fact: “[P]er the protocol, those patients should have been excluded because they were unblinded, correct? . . . Yes.”[103]

Internal Forest documents confirm these patients were, in fact, considered unblinded by the Forest scientists and statisticians working on MD-18.  For example, Ms. Barton sent an email to Drs. Tiseo and Flicker on December 6, 2000, inquiring about whether MD-18 would need to have additional patients enrolled due to the fact “the study drug was unblinded.”[104] In another email, dated August 10, 2001, Jane Wu, a biostatistician working on MD-18, explained they needed to generate tables “excluding the 9 patients who were unblinded at the beginning of the study.”[105] In another email, dated April 5, 2002, Julie Kilbane was finalizing the submission of MD-18 and sent an email explaining that “[s]ome of the supplies were unblinded for this study[.]”[106] Within Forest, there was no ambiguity about whether these patients were actually “unblinded.”

C. Forest Knowingly Misled the FDA about the Nature of the Unblinding by Using, As Forest Regulatory Affairs Manager Put It, “Masterful Euphemisms” to “Protect Medical and Marketing”

After correcting the packaging error to prevent further “automatic” unblinding, Forest debated whether to notify the FDA of the problem.  Dr. Tiseo drafted an initial version of a letter to send to the FDA.[107]  Dr. Flicker reviewed it and advised not sending any letter, but, if Forest did send a letter, he advised giving considerably less detail.[108]  After incorporating Dr. Flicker’s comments, Dr. Tiseo circulated a draft version of the letter to various Forest executives and regulatory personnel, including Lawrence Olanoff, Ivan Gergel, Amy Rubin, Tracey Varner, Julie Kilbane, and Dr. Flicker.[109]  The draft letter stated that the dispensed medication could have “unblinded the study.”[110]  Dr. Tiseo solicited comments from the group.[111]

Amy Rubin, who worked in Regulatory Affairs, edited the letter, changing the language from stating that the dispensing error could have “unblinded the study” to stating the error had the “potential to cause patient bias.”[112]  Ms. Rubin’s edit drew criticism from Dr. Flicker who felt Ms. Rubin’s edits were not up front: “Altho ‘potential to cause bias’ is a masterful stroke of euphemism, I would be a little more up front about the fact that the integrity of the blind was unmistakenly violated.”[113]  This criticism, however, did not prompt Ms. Rubin to correct the language of the letter.  Instead, Ms. Rubin responded: “Thanks for the compliement [sic].  Part of my job is to create ‘masterful’ euphemisms to protect Medical and Marketing.”[114]  For Ms. Rubin, misleading the FDA was not only acceptable, it was part of her job.[115]

And, she did her job well.  The letter ultimately sent to the FDA on March 20, 2000, contained the misleading “masterful euphemism” language.[116]  The letter did not disclose that the patients dispensed the pink pills were “automatically unblinded” as Dr. Tiseo stated to the investigators or that, as Dr. Flicker noted, the integrity of the blind was “unmistakenly violated.”  The smokescreen was up.

As part of the MDL litigation, Plaintiffs deposed Dr. Thomas Laughren, the former Director of Psychiatric Drug Products in the Division of Neuropharmacological Drug Products at the FDA.  He personally reviewed the final study report for MD-18 while at the FDA and, ultimately, was the man at the FDA who approved Lexapro for use in adolescents in 2009.[117]  Dr. Laughren departed the FDA in 2013 and, within months, was working as a testifying expert for various pharmaceutical companies, including Forest.[118]  In fact, Dr. Laughren was hired as a testifying expert for Forest to provide opinions about the pediatric efficacy of Celexa and Lexapro and whether the drugs can increase the risk of suicidal behavior in children.[119] Notwithstanding Dr. Laughren’s unseemly transition from regulating Forest at the FDA to working for Forest, he was shown the euphemism emails and other documents, and he took offense to Forest’s conduct:

  1. [D]oes it concern you that the clinical medical director at the time, Dr. Flicker, believes that a letter that is being proposed to the FDA contains “a masterful stroke of euphemism”?
  2. Yeah, no, that’s – that’s concerning, I would say. . . .
  1. Does it concern you that an employee for Forest whose job it is to interact with the FDA states that it’s part of her job to “create masterful euphemisms to protect medical and marketing”?
  2. It — it is objectionable. I mean, my — my expectation of — of companies is that they will be, you know, completely transparent with — with the FDA about what happened in the conduct of a trial.
  1. Does it concern you that Ms. Rubin, whose job it was to interact with the FDA, believes that it’s her job to “create masterful euphemisms to protect medical and marketing”? . . .
  2. What — what concerns me is — is that — you know, what was represented to FDA was not precisely what happened.[120]

D. Despite Misrepresenting the Unblinding to the FDA, Forest Promised to Exclude the Data from the Patients from Its Primary Efficacy Analysis

While the March 20, 2000 letter to the FDA misrepresented the nature of the unblinding to the FDA, it also stated:  “For reporting purposes, the primary efficacy analysis will exclude the eight potentially unblinded patients, with a secondary analysis including also to be conducted.”[121]  This sentence was added by Dr. Flicker to the original draft of the letter.[122]  Dr. Flicker, consistent with his view that the integrity of the blind was unmistakenly violated, knew the data from these patients was corrupted.  So, Forest promised the FDA, consistent with the express wording of the MD-18 study protocol, that the unblinded patients would not be counted in the primary efficacy analysis and that, instead, “[a] full complement of 160 patients will be enrolled under standard double-blind conditions.”[123] Dr. Flicker acknowledged: “[Y]ou were suggesting that the nine patients subject to the dispensing error were not standardly double-blinded, correct?  . . .  I think it does suggest that.”[124]  Importantly, this promise to exclude the unblinded patients from the primary efficacy analysis was made before Forest knew the results of the primary efficacy endpoint turned on the inclusion of those patients.

E. Forest Reneged on its Promise to Exclude the Unblinded Patients from the Primary Efficacy Results and, Again, Misrepresented the Unblinding in MD-18’s Final Study Report

The protocol for MD-18 only required the enrollment of 160 patients.[125] However, accounting for the unblinded patients, the study would need to randomize at least 169 patients total so that Forest would have a full-complement of patients under standard double-blind conditions. Forest ultimately randomized 174 patients into the study, including nine of the unblinded patients.[126]  So, for purposes of powering the study with sufficient patients, MD-18 did not require the data from the unbinded patients.[127]

After the MD-18 data was collected, however, Forest reneged on its promise to exclude the unblinded patients from the primary efficacy analysis.  Without any consultation with the FDA, Forest slipped the unblinded patients into the primary efficacy analysis—combining the data from the unblinded patients with the blinded cohort—and prepared a secondary “post-hoc” analysis excluding the patients and put it in an appendix.[128]  Including these unblinded patients into the results was “substantial.”[129]  With the unblinded patients in the study, the primary endpoint reached statistical significance, but with the unblinded patients excluded—as Forest promised the FDA—all primary and secondary endpoints were negative.  Had Forest done what it promised, the study would have been negative.  In fact, Forest’s corporate representative conceded that MD-18 is a negative study when the unblinded patients are excluded.[130]  So did Dr. Heydorn, the primary author of the MD-18 study report:

  1. By excluding these nine patients, the P-value went from a statistically significant .038 to a statistically insignificant .052 on the CDRS-R rating scale after 8 weeks, correct?
  2. Yes.
  1. So, in other words, this P-value shows citalopram versus placebo was negative for the primary outcome measure for MD-18, right?
  2. Yes.[131]

Indeed, Dr. Heydorn admitted that, if these patients were unblinded, the fact that the study was negative without them “undermine[s] the assertions that Study 18’s outcome was positive for showing Celexa significantly improved major depression disorder in children and adolescents[.]”[132] When Dr. Heydorn was shown the internal documents demonstrating that the integrity of the blind was unmistakenly violated, he conceded Forest was not honest with the FDA, that Forest misled people about the results of MD-18, and that he would have written the study report differently:

  1. Do you have any regrets about your involvement with the CIT-MD-18 based on what I’ve shown you today?
  2. I wish we had done things a little differently.
  1. Like what?
  2. I wish I had known for certain whether the patients, those nine patients were unblinded, but obviously I don’t know. You showed me a lot of documents today suggesting that people knew the patients were unblinded. I don’t know for a fact that they knew that. All I know is what they wrote on the paper. I wish I was aware of the correspondence with the FDA.
  1. Do you think, based on what I’ve shown you today, that Forest misled anyone about the results of MD-18?
  2. It probably should have been more forthcoming.

  1. Would you have changed anything in the final study report?
  2. If I were the only one involved in writing it, I probably would have written it somewhat differently.[133]

When Dr. Heydorn’s testimony was presented to Dr. Laughren, he testified:

  1. It appears based on Dr. Heydorn’s testimony, he did not believe that the final study report was fully up front or forthcoming with the FDA; isn’t that true?
  2. That’s what he’s saying.
  1. And he’s the man who actually was responsible for the final study report for Study MD-18, right?
  2. He appears to have been, yes.
  1. Does it concern you that Dr. Heydorn, who was a former FDA employee himself, thinks that Forest was not as forthcoming as it should have been with the FDA about its representation of the results from MD-18?
  2. Yes.[134]

The original draft of the MD-18 study report was prepared by a company called PharmaNet, a contract research organization.[135]  Before the first draft of the report was prepared, Dr. Flicker, Dr. Heydorn, and two biostatisticians from Forest met with PharmaNet to discuss how the report should be prepared.[136]  The notes of the meeting illustrate Forest’s general strategy in dealing with the unblinded patients:

Dosing error – some citalopram table[t]s were not blinded. The 9 patients who received unblinded medication were included in the main analyses; a secondary “Post-hoc analysis of the ITT subpopulation” was done. Refer to these analyses briefly in methods and results and reference the reader to the appendix table.[137]

Thus, from the outset, Forest intended to bury the impact of the unblinded data by referring “to these analyses briefly” and referencing “the reader to the appendix table” on page 244 (of 2,135).  It is also worth noting that, even here, in this meeting, Forest was once again stating that the drugs “were not blinded” and that the 9 patients “received unblinded medication.”  As shown below, this clear admission of unblinding was deliberately removed from the final study report sent to the FDA.

In the final study report for MD-18, there are four references to the unblinded patients and all of them are misleading or factually false.  The first reference is in a section of the Study Report titled “Blinding” where it states:

Because of a drug packaging error, the citalopram or placebo tablets initially dispensed to 9 patients at 3 study centers were distinguishable in color, although otherwise blinded (see Section 7.0). When this error was identified at the beginning of the study period, all study medication shipments were replaced in full with tablets of identical color to remove any potential for unblinding.[138]

This paragraph is riddled with inaccuracies and misstatements.  First, the placebo tablets initially dispensed were not distinguishable in color—only the Celexa group received the pink pills, which is why the investigators were unblinded. If both the placebo and Celexa pills had been pink, then the investigators would not necessarily have known which patients were assigned to each group.  Second, they were not just distinguishable in color—the pink pills were Forest-stamped, dose-stamped, commercial Celexa tablets.  The failure of Forest to disclose that the drug dispensed was commercial branded Celexa is misleading in the extreme.  Third, when this error was identified, the medication for those patients not yet randomized was replaced, but for the nine patients already in the study, Forest did not replace their medications.[139]  Forest instructed each site to continue using the multicolor pills for those patients already randomized to placebo.[140]  That means those Celexa patients received white pills for the screening period, pink pills for the first four weeks, and then white pills for the last four weeks.  This paragraph falsely stated that these patients’ medication was replaced “to remove any potential for unblinding” and this is simply not true.

Notably, in the original draft of the MD-18 study report prepared by PharmaNet, a section of the study report contained the language from the original protocol, specifying that “[a]ny patient for whom the blind had been broken was to be immediately discontinued from the study and no further efficacy evaluations were to be performed.”[141]  Dr. Flicker, however, crossed this language out and inserted the language that ultimately made its way into the final study report:

The second reference to the dispensing error is in the section titled “Changes in the Conduct of the Study and Planned Analysis” and it reads:

Nine patients (Patients 105, 113, 114, 505, 506, 507, 509, 513, and 514) were mistakenly dispensed 1 week of medication with potentially unblinding information (tablets had an incorrect color coating). Therefore, in addition to the analysis specified in Section 6.4.1 for the primary efficacy parameter, a post-hoc analysis was performed on an ITT subpopulation that excluded these 9 patients.[142]

This paragraph is also misleading and factually incorrect.  First, these nine patients were not dispensed one week of medication with potentially unblinding information.  These patients received unblinded drug for the first four weeks, not just one week.[143] When this point was shown to Dr. Heydorn, he admitted the statement about one week was not true: “A. It does say one week of medication, yes. . . . Q. So that’s not actually true, right, with respect to patients 113 and 513, correct? . . . It would appear not to be true, yes.”[144]  And, since the investigators were unblinded, the patients were technically unblinded for the entire study.  Second, once again, Forest stated there was an incorrect color coating, even though the pink pills were actually Forest-stamped, dose-stamped, commercial Celexa tablets.  Third, Forest stated that it was providing “a post-hoc analysis” excluding these nine patients.  But, again, this directly contradicts the letter Forest sent to the FDA when the unblinding occurred: “For reporting purposes, the primary efficacy analysis will exclude the eight potentially unblinded patients, with a secondary analysis including them also to be conducted.”[145]

Finally, the phrase “potentially unblinding information” is deeply misleading.  Ironically, this was Dr. Flicker’s phraseology—the same person who, back in March 2000, characterized “potential to cause bias” as a “masterful stroke of euphemism” and felt that the “integrity of the blind was unmistakenly violated.”[146]   In the original draft of the MD-18 study report, it stated: “Nine patients . . .  accidently received 1 week of unblinded study drug treatment[.]”[147]  Note, there was no “potential” or uncertainty about whether the patients received unblinded study drug treatment.  However, in November 2001, when Dr. Flicker edited the first draft of the report, he crossed out this language and added the “potentially” language:[148]

So, Dr. Flicker stated that the “integrity was unmistakenly violated” in 2000, and then, a year later, after he learned that the unblinded patients were needed to obtain a positive result on the primary endpoint, he characterized it as “potentially unblinding information.”

The third reference to the unblinding is in the section discussing the primary efficacy endpoint.  It reads:

Appendix Table 6 presents the results from the LOCF analysis for the change from baseline to Week 8 excluding data from the 9 patients for whom the study blind was potentially compromised (see Section 5.3.4). The results from the Week 8 LOCF analysis comparing the mean change from baseline in CDRS-R in the citalopram and placebo groups was not substantially affected by the exclusion of those patients; the LSM difference decreased from 4.6 to 4.3 and the p-value increased from 0.038 to 0.052.[149]

This is also misleading. And, once again, this is the handiwork of Dr. Flicker.  The original draft of the study report stated: “Appendix Table 6 presents the results from the LOCL analysis for the change of baseline t week 8 excluding data from the 9 patients . . . who accidently received 1 week of unblinded study drug treatment[.]”[150]  Dr. Flicker crossed out this language and crafted some masterful euphemisms of his own:[151]

What makes this paragraph so misleading—aside from suggesting these patients were not actually unblinded—is that Dr. Flicker stated that the exclusion of the unblinded patients did not substantially affect the results of the study.  But that is just not true.  Excluding the unblinded patients makes the primary endpoint no longer statistically significant, i.e., negative.[152]  It changes the entire result of the endpoint and, by extension, the study.  Dr. Heydorn testified:

  1. So with the dispensing error patients excluded from the MD-18 primary efficacy outcome measure, Celexa failed to significantly outperform placebo in treating pediatric depression, right?
  2. That appears to be the case.
  1. That would be an important substantial difference, wouldn’t it?
  2. Yes.[153]

The final reference to the unblinding was in the section of the report titled, “Validity.”[154]  It reads:

The study was designed to provide a valid, prospectively randomized, double-blind comparison of the treatment effects of citalopram and placebo. A medication packaging error partially compromised the study blind for 9 of the 174 patients. Post-hoc analysis excluding these patients supported the results from the intent-to-treat analysis.[155]

This section of the report was also drafted by Dr. Flicker.[156]  And, like the previous sections, it misstates the effect of excluding the unblinded patients from the trial on the overall results.  Thus, all the sections in the final study report addressing the unblinding issue were drafted by Dr. Flicker and none of them state, as he previously stated in his email, that the integrity of the blind was unmistakenly violated.  The report was deliberately misleading or, at least in Dr. Flicker’s own words, not up front.

F. The FDA Never Fully Considered the Unblinding Issue and a Reasonable Regulator at the FDA Could Review this New Information and Conclude Study MD-18 Was Negative

Forest submitted the MD-18 Study Report to the FDA as part of an application seeking a pediatric indication for Celexa.  Ultimately, the FDA denied the application, stating there was insufficient evidence that Celexa was effective in treating pediatric depression.[157]  A careful review of the FDA’s analysis of MD-18, however, reveals that the FDA was misled about the unblinding situation and, ultimately, the results of the study.

MD-18 was reviewed by Dr. Laughren and Dr. Earl Hearst.[158]  Dr. Laughren was the Team Leader of Psychiatric Drug Products and Dr. Hearst was the primary medical reviewer.[159] Normally, clinical trials are reviewed by more than one medical reviewer and the FDA conducts a statistics review, designed to verify the statistics presented by the drug sponsor.[160]  Dr. Laughren explained that “[t]he — the statistical review would likely go into more detail on the — on the analysis plan and whether or not it was followed in — in conducting the analysis.”[161]  However, because the FDA and Forest understood that Celexa would not be approved for children due to the negative result of Study 94404, the FDA determined “there was no need for a statistics review of the efficacy data.”[162]  Instead, the FDA only did a medical review of Study 94404 and MD-18.

Dr. Hearst’s primary medical review of MD-18 concluded, based on the information in the final study report, that MD-18 was positive.[163]  Regarding the unblinding issue, Dr. Hearst copied and pasted the text from the final study report—the report Dr. Heydorn conceded he would have written differently had he known about the unblinding issue.[164]  Dr. Hearst copied verbatim: “[b]ecause of a drug packaging error, the citalopram or placebo tablets initially dispensed to 9 patients at 3 study centers were distinguishable in color, although otherwise blinded.”[165]  Indeed, all but two words of Dr. Hearst’s review of MD-18 consists of sections copied and pasted from the final study report, suggesting the FDA relied heavily on the accuracy of the report.  And, by copying and pasting from the study report, Dr. Hearst parroted Forest’s assertion that the data from these unblinded patients was not actually unblinded.  Dr. Laughren acknowledged that Dr. Hearst appeared to have copied and pasted from the final study report and conceded this was not the approach he endorsed:

  1. And that is a verbatim copy and paste which was in Dr. Hearst’s medical review, correct?
  2. Yes. . .That — that does look like it’s – it’s identical language.
  1. Now, Doctor, in the course of your work at the FDA, do you recall copying and pasting language from a final study report into your medical review?
  2. No, I — I — I did not do that.
  1. Why not?
  2. Because I preferred to reach my own conclusions.[166]

Dr. Laughren also prepared a memorandum, which included a review of MD-18.[167]  Although Dr. Laughren advised against a pediatric indication for Celexa, he stated in reference to MD-18 that “I agree with Dr. Hearst that this is a positive study in support of the efficacy in pediatric MDD.”[168] With regard to the unblinding issue, he remarked that “[t]here was a packaging error resulting in tablets being distinguishable for drug and placebo for 9 patients (although still blinded).”[169]  When asked about this sentence, Dr. Laughren testified that, based on the information from the final study report for MD-18, it was his understanding that the patients received different color pills but were still blinded:

  1. Okay. Now, in that sentence, before that, you said: “There was a packaging error in tablets being distinguishable for drug and placebo for nine patients, although still blinded.” It was your understanding that the patients, despite getting a different color tablet, were still blinded, correct?
  2. I – I’m assuming that I made that statement based on something that I had seen in — in the supplement.
  1. Okay. So it was your understanding that the patients, despite receiving different color tablets, were still blinded, correct?
  2. Well, that — that was — that was my assumption, correct.
  1. If in fact the patients were unmistakenly unblinded, that is not what you understood at the time that you wrote this memorandum, correct?
  2. I — I — again, this goes back almost 15 years. I’m not sure what my state of mind was at the time that I — that I wrote this memo. But my belief was based on what I’ve written here is that the patients were blinded.[170]

Thus, in the absence of clear statements such as “the blind was unmistakenly violated,” Dr. Laughren believed the study report’s assertion that the patients were not really unblinded and, thus, their inclusion in the primary endpoint analysis was not a cause for concern.  After showing Dr. Laughren the internal documents where numerous Forest employees stated, in no uncertain terms, that the nine patients were unblinded, Dr. Laughren agreed that the final study report for MD-18 misrepresented what happened with regard to the unblinding:

  1. Now, we reviewed the final study report for MD-18. Nowhere in that study report that we reviewed, the portions that we looked at, did it state that the integrity of the blind was unmistakenly violated, did it?
  2. No.
  1. In fact, the final study report stated that they were otherwise blinded, didn’t it?
  2. It — it suggests that there was a potential for unblinding, but didn’t acknowledge that — that the investigators at least, if they received — if they noticed that the tablets had the — you know, the name “Celexa” on them and were commercial tablets, that the investigators at least would have — would have been unblinded with regard to those patients.[171]

Dr. Laughren also testified that the information presented to him during his deposition was “new information” that he did not consider when he reviewed MD-18 at the FDA:

  1. And I’ve also shown you some documents which suggest that Forest didn’t properly disclose that fact to the FDA in its submissions, correct?
  2. It — it certainly would have been my preference that — that Forest be more transparent with FDA about the issue of unblinding. . . .

  1. Now, considering that they weren’t transparent about that issue, do you think – and also in consideration of the fact that Study MD-18 never had a statistical analysis of the efficacy data, do you think that it would be appropriate for the FDA to take another look at this data just to make sure that in fact Study 18 was — was positive as Forest has represented?
  2. It — it isn’t my judgment at this point. . . . So, I mean I — that – that’s for FDA to decide at this point. . . . Whether or not FDA — and I also told you that, in retrospect, I would have had a statistical review done on — on 18.
    . . .
    And it’s – it’s up to FDA to decide whether or not, you know, based on this — on this, you know, new information, which I think is probably new information from FDA because I wasn’t aware of it at the time. But it’s not my call.[172]

According to Dr. Laughren, the information about the unblinding of the patients constituted “new information” that was not available to him—and was in fact misrepresented to him—while at the FDA.[173]  And, even though it was no longer his “call,” Dr. Laughren agreed that a reasonable regulatory person at the FDA could review this new information and conclude that MD-18 was negative:

  1. Do you agree, though, Doctor, that a reasonable regulatory person at the FDA could come to a different conclusion about the positive results of MD-18?
  2. It — this is always a matter of judgment. So the answer would be, yes, different people looking at the same dataset can reach a different conclusion.[174]

G. Forest Also Misled the FDA about the Results of the Secondary Endpoints

Forest did not limit its deception to the unblinding issue—Forest also misled the FDA about the secondary endpoints in the study.  There were four secondary endpoints: (1) GGI-I at 8 weeks; (2) change from baseline in CGI-S at 8 weeks; (3) change from Baseline in CGAS at 8 Weeks; and (4) change from Baseline in K-SADS-P Depression Module at 8 Weeks.[175]  There is no dispute that all of the secondary endpoints in the study were negative at 8 weeks, meaning none of the secondary measures were statistically significant (p<0.05).[176]  Forest admits they were negative.[177]  And yet, in the final study report for MD-18, under the section titled “Efficacy Conclusions” Forest stated:

Significant differences (p<0.05), indicative of greater improvement in citalopram patients than placebo patients, were also observed on the CGI-I, CGI-S, and CGAS.  Statistically significant effects were not found as consistently across study timepoints for the secondary efficacy parameters as for the primary efficacy parameter, but numerically greater improvement in the citalopram group was observed on every efficacy parameter at every clinic visit in both the LOCF and OC analyses.[178]

This is misleading because it suggests that the CGI-I, CGI-S, and CGAS were statistically significant “indicative of greater improvement in citalopram” when, in fact, they were all negative at week 8, i.e., the pre-specified secondary endpoints.  Internal documents indicate that this was a deliberate strategy.  Specifically, when Forest contracted with PharmaNet to prepare the first draft of the MD-18 study report, they had a conference on October 4, 2001, where Dr. Heydorn, Dr. Flicker, Dr. Jin, and Dr. Wu from Forest attended.[179]  Notes from the conference call indicate Forest knew the secondary endpoints were negative, but wanted to spin the data by focusing on earlier time points in the study when the secondary endpoints were positive:

For secondary efficacy measures – no significant difference at the week 8 LOCF analysis. The[re] are some significant findings early on in treatment. Forest looking at individual patient listings to see if there are any clues as to why week 8 findings were not positive. For now, emphasize the positive findings at earlier time points for the secondary efficacy variables.[180]

Then, when PharmaNet prepared the first draft of the study report, it stated in the section titled “Efficacy Conclusions” that:

All other efficacy parameters showed a consistent numerical trend in favor of citalopram treatment, but failed to reach statistical significance at week 8.  Except for the CGI-I responder score, all other parameters with evaluations at week 6 reached statistical significance in favor of citalopram treatment at this timepoint.  The by-visit evaluations for these parameters show a marked improvement in the placebo scores at the week 8-timepoint, suggesting a placebo effect.  No explanation is currently available for this observation.  This large placebo effect may be, in part, responsible for the lack of statistical significance in favor of citalopram at week 8.[181]

However, Dr. Flicker crossed out this language and handwrote the language that, for the most part, ended up in the final study report.[182]  Notably, Dr. Flicker also attempted to change the definitions of the secondary efficacy parameters to make them incorporate earlier time points, as opposed to week 8.[183]  For example, in the original protocol for MD-18, which Dr. Flicker signed, it stated that “the endpoints for the secondary objectives are the CGI-Improvement score, and change from baseline in the CGI-Severity score, K-SADS-P (depression module) score and CGAS score at Week 8.[184]”  And, PharmaNet faithfully described these objectives in terms of week 8 in its original draft.[185]  However, Dr. Flicker crossed out “week 8” in his editing, suggesting that the endpoint was not at week 8, but at any time period during the study:[186]

Then, in the section describing the Secondary efficacy parameters, once again Dr. Flicker crossed out any reference to week 8:[187]

This elimination of “week 8” from these sections indicates that Dr. Flicker was deliberately attempting to redefine the efficacy parameters so that Forest’s focus on earlier time points, i.e., not at week 8, would appear to be consistent with the study protocol.

The impact of this deception on the FDA’s review of MD-18 is striking.  Normally, as Dr. Laughren explained, the FDA does not pay much heed to the words used in the final study report: “often when a clinical reviewer gets an application, they often go right to the data rather than even reading the summary, because they don’t want to be influenced by — by, you know, the company’s spin on the data.  So they just go right to the datasets and the tables and look at the data.”[188]  That, however, did not happen here.  Instead, Dr. Hearst, who conducted the primary medical review of MD-18 lifted, verbatim, the section of the final study report dealing with the secondary endpoints into his medical review:

Table 3 – Comparison of MD-18 Study Report & Dr. Heart Medical Review

Exh. 9, MD-18 Final Study Report, pg. 72 Exh. 22, Hearst Medical Review, pg. 11
Significant differences (p<0.05), indicative of greater improvement in citalopram patients than placebo patients, were also observed on the CGI-I, CGI-S, and CGAS. Statistically significant effects were not found as consistently across study timepoints for the secondary efficacy parameters as for the primary efficacy parameter, but numerically greater improvement in the citalopram group was observed on every efficacy parameter at every clinic visit in both the LOCF and OC analyses. Significant differences (p<0.05), indicative of greater improvement in citalopram patients than placebo patients, were also observed on the CGI-I, CGI-S, and CGAS. Statistically significant effects were not found as consistently across study timepoints for the secondary efficacy parameters as for the primary efficacy parameter, but numerically greater improvement in the citalopram group was observed on every efficacy parameter at every clinic visit in both the LOCF and OC analyses.

This is Dr. Hearst’s only discussion of the results of the secondary endpoints—the only medical reviewer of MD-18—and it was lifted, verbatim, from Forest’s study report.  When Dr. Laughren was shown this data, he admitted that Forest’s spin on the secondary endpoints had made its way into the FDA official medical review:

  1. Putting Dr. Hearst aside, I’m talking about Forest, we saw that they had a conference where they said they were going to emphasize this.
  2. Yes. Yes. No, it’s — it is consistent with — with that view of focusing on the positive and not giving a complete picture.
  1. And it appears that that spin that Forest put into the final study report made it into Dr. Hearst’s report, correct?
  2. It — it appears to have, yes.[189]

In Dr. Laughren’s review of MD-18—the only other person within the FDA to review MD-18—his discussion of the secondary endpoints was even more inaccurate than Dr. Hearst’s.  He stated: “Results also significantly favored citalopram over placebo on most secondary outcomes” even though every secondary endpoint was negative.[190]  Dr. Laughren could not remember what he was thinking when he wrote the statement:

  1. Now, on page 3, just above the paragraph that says “comment,” there is a sentence that reads: “Results also significantly favored citalopram over placebo on most secondary outcomes.” Do you see that?
  2. Yes.
  1. Now, you didn’t state there that all the prespecified secondary endpoints were negative at week 8, right?
  2. Correct.
  1. You’re referring here, I assume, to the earlier time points when there were statistically significant results in the secondary endpoints, correct?
  2. I — again, I don’t — this was written a long time ago. I don’t recall what would have been in my mind at the time that I wrote this, but it – you’re correct in saying that it doesn’t — it doesn’t emphasize the fact that the eight-week results were all negative on the secondary endpoints.
  1. Now, I know you don’t recall this, but is it possible that when you were drafting this memo, you looked at the final study report, looked at Dr. Hearst, who you relied upon, and thought, oh, most of the secondary endpoints must have been positive?
  2. I — I would — I would have to speculate about what — what I was looking at at the time when I wrote this, and I — I — I prefer not to do that. I just — I don’t know.
  1. Okay. Would you agree with me, though, that it would be accurate to say all the protocol-specified secondary endpoints for Study MD-18 were negative at week 8?
  2. That is — that appears to be correct, yes.
  1. And would you agree with me that — that you don’t state that in your memo?
  2. I — I do not state that in my memo.
  1. And you would agree with me from what we’ve seen in Dr. Hearst’s clinical review, he did not state that either.
  2. He did not appear — appear to do that either.[191]

Thus, from the outset, Forest had an objective—avoid disclosing that all the secondary endpoints were negative and, instead, focus on the secondary measures that reached statistical significance at various earlier time points.  This “spin” was successful.  Not only did the primary medical reviewer at the FDA, Dr. Hearst, copy and paste that spin into his review, but his supervisor, Dr. Laughren went even further by stating that most of the secondary endpoints supported efficacy even though they were all negative.  Like a cascade, starting with Forest’s deliberate decision of obscure the secondary endpoints, the deception made its way into the FDA’s own “independent” reviews.

Part I | Part II | Part III | Part IV | Part V | Part VI

_______________________________________

[45]  Exh. 35, Excerpts of Study 94404 Rpt. at *1.

[46] Id. at *11 (pg. 47 of 345).

[47] Id. at *8-10 (pgs. 42-44 of 345).

[48] Id. at *12-23 (pgs. 58-69 of 345) (data listed in text).

[49] Exh. 36, Letter from Lundbeck to Forest (w/attachment), at 1; Exh. 37, Email re. 94404 Headline results at 1-2 (“94 404 citalopram vs placebo in the treatment of adolescent depression have been unblinded and unfortunately with a negative result. It was not possible to detect a significant difference between the two treatment groups.”).

[50] See Exh. 3, Criminal Information ¶ 70 (“FOREST PHARMACEUTICALS aggressively publicized and promoted the results from the positive Forest study, while at the same time FOREST PHARMACEUTICALS did not publicize or disclose the results of the negative study to persons outside the FDA or the Danish company which sponsored the negative study. As a result, doctors and psychiatrists received incomplete and misleading information concerning all available known data pertaining to the efficacy of using Celexa to treat depression in children and adolescents.”).

[51] Exh. 38, Excerpts of 2007 Depo. of W. Heydorn at 77:23-80:5.

[52] Exh. 39, Email re. Publications at 1-2 (“I just wanted to check on the status for the Wagner pediatric manuscript . . . investigators in the Lundbeck sponsored study seem eager to submit a manuscript on their study (they are working on it – we have not yet seen any draft) and I wanted to make sure that the positive data are in the public domain before their negative data get out.”)

[53] Exh. 40, Barry Meier, Medicine’s Data Gao – Journals in a Quandary; A Medical Journal Quandary:  How to Report on Drug Trials, NY Times (June 21, 2004).

[54] Exh. 41, Press Release, Forest Laboratories, Inc., Forest Discusses Disclosure of Citalopram Clinical Trial Data in Children and Adolescents (June 24, 2004).

[55] Exh. 42, Anne-Liis von Knorring, et al, A Randomized, Double-blind, Placebo-controlled Study of Citalopram in Adolescents with Major Depressive Disorder, 26 J. Clin. Psychopharmacology 3, 311-15 (2006).

[56] Exh. 9, Excerpts of Study MD-18 Rpt. at pg. 38.

[57] Id. at pg.1.

[58] Id. at pgs. 49-50.

591] Id. at pg. 111.

[60] Id. at pgs. 101-104, 111, 244.

[61] Exh. 43, Excerpts of Study MD-18 Protocol at pg. 328.

[62] Exh. 38, Excerpts of 2007 Depo. of W. Heydorn at 42:13-44:13.

[63] Exh. 11, 2016 Depo. of W. Heydorn at 107:13-107:21.

[64] Exh. 30, FDA, supra note 33, at 4 (emphasis added).

[65] See Exh. 43, Excerpts of Study MD-18 Protocol at pg. 334; Exh. 12, 2016 Depo. of S. Closter (Forest’s Rule 30(b)(6) Corporate Representative) at 168:1-169:12; 170:5-14; Exh. 11, 2016 Depo. of W. Heydorn at 29:14-32:2.

[66] Exh. 43, Excerpts of Study MD-18 Protocol at pg. 334; Exh. 11, 2016 Depo. of W. Heydorn at 29:14-32:2.

[67] Exh. 11, 2016 Depo. of W. Heydorn at 29:14-32:2; Exh. 12, 2016 Depo. of S. Closter (Forest’s Rule 30(b)(6) Corporate Representative) at 168:1-169:12; 170:5-14.

[68] Exh. 12, 2016 Depo. of S. Closter (Forest’s Rule 30(b)(6) Corporate Representative) at 168:1-169:12; 170:5-14.

[69] Exh. 38, Excerpts of 2007 Depo. of W. Heydorn at 42:13-44:13; Exh. 11, 2016 Depo. of W. Heydorn at 29:14-32:2.

[70] Exh. 43, Excerpts of Study MD-18 Protocol at pg. 326.

[71] Id. at pgs. 323-26.

[72] Id.

[73] Id.

[74] Id.

[75] Id. at 318 (“Patients must have a Children’s Depression Rating Scale-Revised (CDRS-R) score of 40 or greater at both the Screening and Baseline visits.”)

[76] Id. at pgs. 323-26.

[77] Id.

[78] Id.

[79] Id.

[80] Id.

[81] Id.

[82] Id.

[83] Id.

[84] Id.

[85] Id.

[86] Id.

[87] Id.

[88] Exh. 13, Draft FDA Letter with C. Flicker Handwritten Comments at 1; accord Exh. 11, 2016 Depo. of W. Heydorn at 197:18-198:14 (verifying that the handwriting belongs to Charles Flicker); Exh. 44, Emails re. Urgent CIT-MD-18 at *1 (When notified of the findings by a site, due to seeing white and pink tablets, all supplies were returned and the 10ct bottles re-packaged with non-trade “white” tablets.” (emphasis added)).

[89] Exh. 13, Draft FDA Letter with C. Flicker Handwritten Comments at 1.

[90] Exh. 15, Memo re. CIT-MD-18 (Deviation Report) at 1-2.

[91] Exh. 13, Draft FDA Letter with C. Flicker Handwritten Comments at 1; accord Exh. 16, Email re. CIT-18 FAX to Investigational sites (w/ attachment) at 1 (“[A] copy of the FAX that went out to all CIT-MD-18 Pediatric Investigational sites this morning is attached. All sites have also been contacted by telephone and given verbal instructions on how to proceed[.]”).

[92] Exh. 16, Email re. CIT-18 FAX to Investigational sites (w/ attachment) at 2.

[93] Id. (emphasis added).

[94] Id. at 4.

[95] Id.

[96] Exh. 45, 2016 Depo. of C. Flicker at 278:24-279:6.

[97] Exh. 11, 2016 Depo. of W. Heydorn at 202:13-19.

[98] Exh. 46, J. Jureidini Expert Report at 5.

[99] Id.

[100] See Exh. 47, J. Glenmullen Expert Rpt. at 24 (discussing how the dispensing process for these unblinded Celexa patients occurred).

[101] Id.

[102] Exh. 43, Excerpts of Study MD-18 Protocol at pg. 328.

[103] Exh. 11, 2016 Depo. of W. Heydorn at 227:5-10, 228:20-24, 244:11-17.

[104] Exh. 48, Email re. CIT-MD-18 Study Drug at 1.

[105] Exh. 20, Emails re. CIT-MD-18 at 1.

[106] Exh. 44, Emails re. Urgent CIT-MD-18 at *2.

[107] Exh. 13, Draft FDA Letter with C. Flicker Handwritten Comments at 1; Exh. 11, 2016 Depo. of W. Heydorn at 197:18-198:14 (verifying that the handwriting belongs to Charles Flicker).

[108] Exh. 13, Draft FDA Letter with C. Flicker Handwritten Comments at 1 (“Reconsider no letter otherwise I recommend much less narrative[.]”).

[109] Exh. 17, Email re. Letter to FDA for CIT-18 (w/attachment) at 1 (“Attached please find the letter that Charlie and I put together for the purpose of informing the FDA of our packaging mishap in the citalopram pediatric study.”).

[110] Id. at 2 (“[D]ue to a clinical supplies packaging error for the above-referenced trial, eight randomized patients at two investigational sites were dispensed medication that could have potentially unblinded the study.”).

[111] Id. at 1 (“please review and send your comments back to me within the next few days.”)

[112] Exh. 18, Email responses re. Letter to FDA for CIT-18 at 1-2.

[113] Id. at 1.

[114] Id.

[115] Id.

[116] Exh. 19, Letter from T. Varner (Forest) to R. Katz (FDA) at 1.

[117] Exh. 8, 2017 Depo. of T. Laughren at 401:24-402:8 (“[Y]ou’re actually the one who ultimately signed off finally on Lexapro’s approval for adolescents, right?  A Yes.”); Exh. 49, Lexapro Approval Letter for Adolescents at 3 (signed by Dr. Laughren).

[118] Exh. 8, 2017 Depo. of T. Laughren at 81:1-83:14.

[119] Id.

[120] Id. at 205:14-21, 206:14-23, 207:21-208:4 (emphasis added).

[121] Exh. 19, Letter from T. Varner (Forest) to R. Katz (FDA) at 1 (emphasis added).

[122] Exh. 13, Draft FDA Letter with C. Flicker Handwritten Comments at 1.

[123] Exh. 19, Letter from T. Varner (Forest) to R. Katz (FDA) at 1.

[124] Exh. 45, 2016 Depo. of C. Flicker at 294:13-23.

[125] Exh. 43, Excerpts of Study MD-18 Protocol at pg. 318 (“The study population will be equally stratified between children (aged 7 to 11) and adolescents (ages 12 to 17).  A total of 160 patients will be randomized to double-blind treatment.”)

[126] Exh. 9, Excerpts of Study MD-18 Rpt. at pg. 64 (“A total of 174 patients received double-blind study drug, of whom 89 received citalopram and 85 received placebo.”

[127] Id. at pg. 62 (“Assuming an effect size (treatment group difference relative to pooled standard deviation) of 0.5, a sample size of 80 patients in each treatment group was used[.]”).

[128] Id. at pg. 63.

[129] Exh. 11, Depo. of W. Heydorn at 88:3-17 (Dr. Heydorn admitting that including the patients made an “important substantial difference” and that those patients were not needed to power the study).

[130] Exh. 12, 2016 Depo. of S. Closter (Forest’s Rule 30(b)(6) Corporate Representative) at 294:10-295:20 (“If they were removed from the study, I understand that the result would have been negative.”).

[131] Exh. 11, 2016 Depo. of W. Heydorn at 86:22-87:9.

[132] Id. at 112:14-112:20.

[133] Id. at 308:16-309:6 (emphasis added).

[134] Exh. 8, 2017 Depo. of T. Laughren at 263:9-264:5 (emphasis added).

[135] Exh. 11, 2016 Depo. of W. Heydorn at 237:2-15.

[136] Id. at 236:15-237:6; Exh. 21, Email re. Notes from conference call Oct 4 (w/attachment) at 1 (“Attached are my notes from the conference call with the CRO on the peds study.”).

[137] Exh. 21, Email re. Notes from conference call Oct 4 (w/attachment) at 2 (emphasis added).

[138] Exh. 9, Excerpts of Study MD-18 Rpt. at pg. 44.

[139] Exh. 16, Email re. CIT-18 FAX to Investigational sites (w/ attachment) at 3 (“Patients already randomized . . . will proceed through the study normally a . . . DO NOT ship their remaining drug back to Forest. Keep all of their drug at your site and continue to use it as you would ordinarily. Return only the units of drug for your non-randomized patients.”).

[140] Id.

[141] Exh. 50, Draft of MD-18 Study Report w/ C. Flicker Comments at 26.

[142] Exh. 9, Excerpts of Study MD-18 Rpt. at pg. 63.

[143] Indeed, investigators were not notified of the problem until Dr. Tiseo sent out the facsimile on March 2, 2000.  See Exh. 16, Email re. CIT-18 FAX to Investigational sites (w/ attachment) at 1.  At that point in time, three patients had already been in the study for over a month, and the rest had been in the study for over two weeks.  Exh. 9, Excerpts of Study MD-18 Rpt. at pg.1214-15, 1235-37 (listing the dates of each unblinded patients’ various assessments). The statement that these patients only received incorrectly colored drug for one week is plainly false.

[144] Exh. 11, 2016 Depo. of W. Heydorn at 176:3-20 (emphasis added).

[145] Exh. 19, Letter from T. Varner (Forest) to R. Katz (FDA) at 1.

[146] Exh. 18, Email responses re. Letter to FDA for CIT-18 at 1 (emphasis added).

[147] Exh. 50, Draft of MD-18 Study Report w/ C. Flicker Comments at 44 (emphasis added).

[148] Id.

[149] Exh. 9, Excerpts of Study MD-18 Rpt. at pg. 70.

[150] Exh. 50, Draft of MD-18 Study Report w/ C. Flicker Comments at 49 (emphasis added).

[151] Id.

[152]  Exh. 12, 2016 Depo. of S. Closter (Forest’s Rule 30(b)(6) Corporate Representative) at 294:10-295:20.

[153] Exh. 11, 2016 Depo. of W. Heydorn at 87:19-88:6.

[154] Exh. 9, Excerpts of Study MD-18 Rpt. at pg. 83.

[155] Id.

[156] Exh. 50, Draft of MD-18 Study Report w/ C. Flicker Comments at 67.

[157] Exh. 51, Letter from R. Katz (FDA) to T. Varner (Forest) at 1-2 (“[A] single positive study is not sufficient, in our view, to support this new claim in pediatric MDD . . . the history of predominantly negative placebo-controlled trials in pediatric MDD argues against the extrapolation of the MDD claim from adults to pediatric patients on the basis of the adult data alone, or even on the basis of one positive study in pediatric patients, along with positive adult data.”).

[158] Exh. 52, T. Laughren, Memo. re. Recommendation for Non-Approval at 1.

[159] Id. (“The primary review of the clinical efficacy and safety data was done by Earl Hearst, M.D., from the clinical group.”).

[160] Exh. 8, 2017 Depo. of T. Laughren at 95:10-96:3.

[161] Id. at 95:19-22.

[162] Note also Laughren Depo. at 94:1-95:6.

[163] Exh. 22, Review and Evaluation of Clinical Data by Dr. Earl Hearst, FDA at 2.

[164] Compare id. at 11 (“Because of a drug packaging error, the citalopram or placebo tablets initially dispensed to 9 patients at 3 study centers were distinguishable in color, although otherwise blinded.”) with Exh. 9, Excerpts of Study MD-18 Rpt. at pg. 44 (“Because of a drug packaging error, the citalopram or placebo tablets initially dispensed to 9 patients at 3 study centers were distinguishable in color, although otherwise blinded[.]”).

[165] Exh. 22, Review and Evaluation of Clinical Data by Dr. Earl Hearst, FDA at 11.

[166] Exh. 8, 2017 Depo. of T. Laughren at 278:9-278:16 (emphasis added).

[167] Exh. 52, T. Laughren, Memo. re. Recommendation for Non-Approval at 1.

[168] Id. at 3.

[169] Id.

[170] Exh. 8, 2017 Depo. T. Laughren at 154:6-155:9.

[171] Id. at 205:4-13.

[172] Id. at 397:9-398:12.

[173] Dr. Laughren also testified that he did not believe the new information would have changed his judgment that MD-18 was a “positive” study, because even though all the secondary endpoints were negative and the p-value for the primary efficacy endpoint went above 0.05 with the unblinded patients excluded, he testified it was “close enough.”  Id. at 147:7-148:11, 168:21-169:5.  This opinion, however, strains credibility.  Back in 2013, before the unblinding issue was unearthed, Dr. Laughren was deposed as an expert for Forest and he specifically testified that exclusion of the unblinded patients rendered MD-18 negative: “Q. [I]f these patients were removed, this would no longer be a positive study? A. That’s correct.”  Exh. 53, Excepts of 2013 Depo. T. Laughren at 301:20-302:2 (emphasis added).  Indeed, Forest and Dr. Heydorn both agree that MD-18, with the unblinded patients excluded, is negative.  Exh. 12, 2016 Depo. of S. Closter (Forest’s Rule 30(b)(6) Corporate Representative) at 294:10-295:20 (“If they were removed from the study, I understand that the result would have been negative.” (emphasis added)); Exh. 11, 2016 Depo. of W. Heydorn at 87:11-87:14 (same).  Dr. Laughren’s “close enough” opinion is an after-the-fact attempt to justify his conclusion that MD-18 was positive—a conclusion that formed the basis of his approval of Lexapro for use in adolescents in 2009.  To admit that the study would be negative while excluding the unblinded patients would force him to concede that he made a mistake in approving Lexapro for use in adolescents.

[174] Exh. 8, 2017 Depo. of T. Laughren at 402:18-403:2 (emphasis added).

[175] Exh. 43, Excerpts of Study MD-18 Protocol at pg. 321, 329.

[176] Exh. 9, Excerpts of Study MD-18 Rpt. at pgs. 101-104.

[177] Exh. 12, 2016 Depo. of S. Closter (Forest’s Rule 30(b)(6) Corporate Representative) at 188:4-20 (confirming, on behalf of Forest that all secondary endpoints were negative).

[178] Exh. 9, Excerpts of Study MD-18 Rpt. at 72.

[179] Exh. 21, Email re. Notes from conference call Oct 4 (w/attachment) at 1.

[180] Id. at 2.

[181] Exh. 50, Draft of MD-18 Study Report w/ C. Flicker Comments at 51-52.

[182] Id.

[183] Id. at 35-36, 38.

[184] Exh. 43, Excerpts of Study MD-18 Protocol at pg. 321, 329.

[185] Exh. 50, Draft of MD-18 Study Report w/ C. Flicker Comments at 35.

[186] Id.

[187] Id. at 38.

[188] Exh. 8, 2017 Depo. of T. Laughren at 68:18-69:4 (emphasis added).

[189] Exh. 8, 2017 Depo. of T. Laughren at 279:6-17 (emphasis added).

[190] Exh. 52, T. Laughren, Memo. re. Recommendation for Non-Approval at 3.

[191] Exh. 8, 2017 Depo. of T. Laughren at 279:21-282:2 (emphasis added).