SSRI RISKS: NEW DATA OR RENEWED SCRUTINY?
By David Healy, MD, Ph.D
Even before Prozac was licensed at the end of 1987 regulators in Germany had raised the issue of suicidality. Prozac was held up for many years in Germany and only finally licensed at the end of 1989 after some apparently unorthodox lobbying of supposedly independent regulatory panel members. In the course of handling various problems including suicidality before and after launch in the US, executives from Lilly had access to the head of the CNS Division of the FDA out of hours, and to FDA Director David Kessler at his home, as well as to other officials involved in vetting the applications for other SSRIs such as Paxil.
Did this access help the cause of patient safety? A lot depends on how one interprets how Lilly and the FDA managed the assessments of Prozac in Germany. In a letter to Lilly on September 9th, 1987, indicating Prozac was approvable, Robert Temple of the FDA wrote:
We require a review of the status of all fluoxetine actions taken or pending before foreign regulatory authorities. Approval actions can be noted, but we ask that you describe in detail any and all actions taken that have been negative, supplying a full explanation of the views of all parties and the resolution of the matter.
In response, on October 26th, 1987, Lilly submitted a series of reports including one from the German regulatory body, the BGA. The first point (1.0) the Germans made was "the drug is not sufficiently tested .. and the therapeutic efficacy claimed for [Prozac] is insufficiently substantiated." The second point (2.0) referred to Prozac's: "unacceptable damaging effects". Following this, under 2.1, the BGA noted "[an] increase in agitating effect occurs earlier than the mood elevating effect and therefore an increased risk of suicide exists." Under 2.2, the BGA noted that during treatment with the drugs "some symptoms of the underlying disease (anxiety, insomnia, agitation) increase, which as adverse effects exceed those which are considered acceptable by medical standards." The letter containing all this was sent to the FDA by Max Talbott, Lilly's regulatory advisor, a former FDA employee himself.
It is not clear what the FDA understood unacceptable damaging effects to mean or whose job it was÷Lilly's or the FDA's÷to make sure the FDA understood. The FDA routinely sends letters to companies prior to approving drugs asking to be fully informed about any decisions made by foreign regulators. But it is clear that the question was asked as to just what unacceptable and damaging meant. There were phone calls between Laughren and Kapit of the FDA and members of Lilly, which led to a letter from Talbott on December 4th stating "In February 1985... the BGA had alluded to 'unacceptable damaging effects'· This phrase was not defined." He went on to say that "the data submitted to the BGA have also been submitted to the FDA, which has reviewed data on more patients using more sophisticated analyses than those submitted to the BGA up to this time."
A memo from Richard Kapit on December 8th 1987 records receipt of this letter, the teleconference that had given rise to it, and the assertion that the
...German authorities never defined or documented the phrases 'severe organ damage' or 'unacceptable damaging effect.' The company asserts that all information made available to the BGA has been made available to the FDA... Conclusion: the BGA comments do not appear to reflect clinical events, since no such events have been reported to the FDA and according to the company we have received all information submitted to the BGA. Recommendation: The comments by the BGA should not affect FDA's conclusion that NDA 18-936 is approvable.
By the end of December that year Prozac had been approved.
Nearly seven years later, having claimed that if he even overheard someone report a side effect of Prozac in a corridor at a party he would report it to the FDA, Nancy Zettler, the attorney for the plaintiffs in the Wesbecker case, faced Leigh Thompson, then the chief scientist at Lilly, with this issue in the course of his deposition in the Wesbecker case:
Q: Do you feel that you have an obligation to report the fact that the German government raised the issue and that it was fully analyzed back in 1984 or '85?
A: I would probably call on my legal colleagues to help me interpret it, but under the current NDA regulation, I think the current NDA regulation does request information on negative findings by other regulators. But what that means has never been tested in court in terms of whether that is just a letter that you receive from the BGA saying we're not going to approve your drug, or whether it means everything that you talk about in working up whatever their questions are. I'm not aware that that's ever been resolved.
Q: So let me make sure I understand. You feel it is your absolute duty to report a conversation that you have overheard in the hallway, but you don't necessarily feel it's your duty to report an issue raised by another regulatory agency and the analyses that were done by the drug company in response to that issue being raised?
A: Under the FDA laws and regulations, I think you've stated it correctly. 
It seems fairly evident what the unacceptable damaging effects were. It is not clear that the FDA sought or received advice from anyone on this point. There is no record of what any of these FDA officials made of the testimony of Joachim Wernicke, a key player at Lilly in steering Prozac through the FDA who, faced with this very point in the Wesbecker trial, acknowledged that a mistake had been made.
Aside from any interpretations of what these exchanges reveal, there are a number of datasets that throw light on what might be said about what Lilly or the FDA can reasonably be said to have done or not done. Following the public controversy about the risk of suicide induction on Prozac, the FDA held a hearing in September 1991 that appeared to exonerate Prozac. At this hearing experts were faced with a selected dataset from Lilly's clinical trial data.
Some years beforehand, in 1985, Lilly had presented a much more comprehensive set of clinical trial data to the German regulators. One explanation for the difference in the datasets is that the 1991 set contained US clinical trial data only. These US trials contained data from a Dr Cohn, which the FDA had previously adjudged to be unsatisfactory.
In this German dataset, Lilly included under the heading of comparator all data from placebo and other antidepressants with which Prozac was being compared. These are the tabulated figures that can be found in the submissions to Germany, but the clinical trial data going with these figures make it clear that Lilly had miscoded as placebo suicidal acts, acts that had not occurred on placebo - occurring either in the run in phase of trials or up to a year after the trial was over.
When this miscoding is taken into account, the resulting correct set of figures is as follows:
These figures were not presented to the FDA panel hearing on Prozac - although they were available to the FDA.
But in addition to these figures, the FDA was at this point sitting on figures for suicidal acts from clinical trials of Zoloft and Paxil, of which panel members were unaware.
The data on Zoloft and Paxil subsequently appeared as follows:
In fact FDA medical reviews of these data, which are publicly available, reveal that Pfizer and SmithKline had followed Lilly's example in miscoding figures for suicidal acts, incorporating under the heading of placebo suicidal acts that had not occurred on placebo. This reorganizing of the data had happened after the public controversy about Prozac had blown up, making it difficult to resist the interpretation that the companies and the regulators were trying to manage a problem. Adjusting for this leads to the following table:
There seem to be three questions for regulators to answer.
This data is now over a decade old.
David Healy MD MRCPsych
. Depositions of Nick Schulz-Solce and Hans Weber in Fentress Vs Eli Lilly.
. Deposition of W Leigh Thompson in Fentress Vs Eli Lilly July 20th, 21st & 22nd (1994).
. Letter to Max Talbott of Lilly from Robert Temple of FDA on September 9th 1987, in a series of correspondences that followed 88 submissions from Lilly about Prozac starting in September 6th 1983. Fentress Exhibit Pz1064 290.
. Letter from MaxTalbott to FDA, Fentress Exhibit Pz 2027, 1643.
. See Fentress trial testimony÷Joachim Wernicke.
. Letter from Max Talbott to the FDA, December 4th 1987, Exhibit 10 in Deposition of N Schulz-Solce in Fentress Vs Eli Lilly.
. Memo from Richard Kapit of FDA, Dec 8th 1987, Exhibit 11 in Deposition of N Schulz-Solce in Fentress Vs Eli Lilly.
. Deposition of W Leigh Thompson in Fentress Vs Eli Lilly July 22nd 1994.
. Testimony of Joachim Wernicke in Fentress Vs Eli Lilly.
 Brickler G (1994). Exhibit 1 in the deposition of G Brickler in Fentress Vs Eli Lilly
 Healy D (2003). Lines of Evidence on SSRIs and Risk of Suicide. Psychotherapy and Psychosomatics 72, 71-79.
to Quit Psychiatric Medication
For tech support, contact webmaster