Healy - One Side of the Background to an Academic Freedom Dispute

One Side of the Background to an Academic Freedom Dispute

by David Healy

 Background

When the guest speaker at the annual meeting of the British Association for Psychopharmacology in July 2000 came up to me at a BAP poster stand, where I was presenting details of a study, and said that I had no right to present research like this, I was shocked.  Even when challenged with the fact that there was other unpublished data, which amply bore out the findings and conclusions being reported, he still insisted that I should not be presenting research like this.  Frighteningly, he went on to make it clear that he had been approached to get involved in a set of legal actions “against me”.   

The study was one that was designed to explore whether antidepressants selective to different neurotransmitter systems have different functional effects.  With colleagues, I recruited twenty senior medical, nursing and administrative staff from the psychiatric unit in which the North Wales University Department of Psychiatry was also based.  The volunteers were randomized to a clinical dose of sertraline, a Selective Serotonin Reuptake Inhibitor (trade name Zoloft) or reboxetine, an agent selective to the norepinephrine system (not available in North America) for two weeks, followed by a two-week discontinuation arm and then randomization to the other agent for two weeks.   

Could they specify functional differences between the two agents?  We had a series of other questions such as whether personality type predicted who would prefer which agent.  The study found that drugs selective to the norepinephrine system increased drive and energy whereas a selective serotonin reuptake inhibitor appeared to produce a mellowing of affect.  Personality measures did predict who would prefer these differential effects and we found that 1/3rd of our subjects preferred the effects of sertraline while 1/3rd preferred those of reboxetine.  Roughly half of our subjects felt better than well on one or other of the two agents.   Those who strongly preferred sertraline did rather poorly on reboxetine and vice versa (Tranter et al 2002).  In the case of two of our volunteers the effects of sertraline were disastrous.  They became acutely and seriously suicidal (Healy 2000a). 

By the time I presented these results, I was aware that our findings were not unique.  More than 15 years beforehand Pfizer had run a study in which healthy female volunteers were randomized to sertraline or placebo.  The study terminated early with all those receiving sertraline discontinuing because of problems in the domain of agitation and apprehension.   

This study had come to my attention owing to an involvement in a medico-legal case – Miller versus Pfizer.    Matthew Miller was a 13-year old boy who following a house move had become disruptive at his new school and had been assessed for possible nervous problems.  There were suggestions that he might be mildly depressed.  He was given sertraline and a week later he had hung himself one night.   

Ten years before my involvement in the Miller case, shortly after fluoxetine and fluvoxamine had been launched on the British market, I had written up a pair of clinical cases in which two individuals had become suicidal within the first weeks of treatment with fluoxetine.  The suicidality had cleared up on discontinuation of treatment but had re-emerged when re-challenged with a further antidepressant active on the serotonin system (Creaney et al 1991).  This had led on to a review of SSRI induced suicidality (Healy 1994), which had triggered a series of medico-legal approaches.  In all instances, I had offered the view that the SSRI involved had not caused the problems that the plaintiffs were claiming.   

This picture changed in 1997 when I was approached on the case of William Forsyth a man who after taking Prozac for 10 days had butchered his wife and then killed himself.  In this instance the drug did seem to me to be involved.  But whereas Lilly had settled a great number of other cases prior to this, this case went to trial and in the process I became aware of an increasing number of documents and data from within Lilly and elsewhere, which reinforced my views that SSRIs could cause problems (Healy 1999a).  The Forsyth case led to involvement in the Miller, Motus and Tobin cases.  Miller and Motus involved sertraline and Pfizer, while Tobin involved paroxetine and Glaxo SmithKline.

Toronto

Towards the end of 1998, I had a first overture about a possible move from Britain to the Center for Addiction and Mental Health (CAMH) in the University of Toronto (formerly the Clarke Institute).  This led to interviews with the search panel in the course of 1999 and a formal university job offer, which in early 2000 I accepted.  The position was a Professor of Psychiatry in the University of Toronto and Head of the Mood and Anxiety Disorders Program within CAMH.  Aspects of the recruitment process and visa applications led to four visits to Toronto during the course of 1999 and 2000.  In a series of lectures, I presented research on the history of the antidepressants (Healy 1997) as well as research indicating that Mental Health Services in general may be doing less well than is commonly portrayed (Healy et al 2001).   In the course of various meetings I made no secret of my involvement in the SSRI controversies, on which I had several current publications (Healy and Savage 1998; Healy 1999a; Healy et al 1999).   

On a personal front involvement in SSRI medico-legal issues made no difference to my willingness to undertake clinical trials of psychotropic agents or preparedness to be a consultant to pharmaceutical companies or to lecture at company sponsored symposia or in other venues.  The problems of possible suicide induction as I saw them were ones that could be handled with appropriate warnings and monitoring.   

In August 2000, I was invited to be a guest speaker at a meeting organized for the end of November 2000 to celebrate the 75th Anniversary of the University Department and the 150th Anniversary of the establishment of Clinical Services in Toronto.  I agreed to talk on the topic of psychopharmacology and the government of the self.  This talk reproduced on this website in essence gives the outlines of a then forthcoming book from Harvard University Press – which overviewed the emergence of psychopharmacology, the development of the field and prospects for the future (Healy 2002).  This seemed appropriate in the context of a meeting called Looking Back, Looking Ahead.  

I was due to give the same lecture the following week in Cornell Medical School as part of the Eric T Carlsson Memorial Grand Rounds in the History of Medicine and the Richardson Research Seminar Series.  Part of the lecture had been given previously at the invitation of AstraZeneca.  The full lecture has subsequently been given in Paris, Minneapolis, Cambridge and elsewhere.   

On the day before the meeting in Toronto, I sat on an interview panel to appoint a neuropsychologist who would work with me on the Mood Disorders Program.  I was consulted about the decor of my office as well as computing support.  I discussed removal expenses and associated issues with David Goldbloom, the Physician in Chief of the CAMH.  I mentioned my SSRI medico legal work to him, which he appeared to view as potential departmental funds.  His only concern appeared to be to get me safely ensconced in the University of Toronto, even earlier than I had planned to get there.  I was less than two weeks away from completing the last formalities in the visa process.  

As is customary these days, the lecture in Toronto was rated for presentation and content by the audience, an audience of over 200 people from across the board in the Mental Health Services.  My lecture was rated highest for content.  The same lecture was very warmly received in Cornell the following week, as it has been elsewhere since.  

But in the hours following the lecture, I was told that David Goldbloom had taken exception to some of the points that had been raised.  The points that concerned him, I was told, were that I’d claimed that Prozac could make people suicidal, that Lilly knew about it (a claim not made) and that we were now treating more patients than ever before (see Healy et al 2001).  I arranged to meet with Goldbloom following a celebratory meal that evening.  He was apoplectic.  There are only three points he told me that any one would ever remember from a lecture – and in this case the points they would remember were that Prozac makes people suicidal, that Lilly knew about it and that high dose antipsychotics can cause brain damage.  A further charge that appeared later was that I had claimed that very large proportions of the clinical literature were now ghost-written. 

The following day I left for New York.  I had a schedule, which involved spending three days in Pfizer’s New York archives, where I wanted to look at their unpublished healthy volunteer studies with sertraline, before lecturing in Cornell.

Following the lecture in Cornell there was also a meal.  At this, the Dean of the Medical School, Bob Michels, asked what had happened in Toronto.  Astonished at the enquiry I outlined the story above.  Michels however appears to have known that the CAMH had taken the step to rescind my job offer as they saw it; breach their contract with me, as I was later to see it.  How had he known? 

I still do not know the answer to this but it turned out that that one of the lecturers at the Toronto meeting, Dr Charles Nemeroff, attending American Foundation for Suicide Prevention council meetings in New York the day after the Toronto meeting had volubly raised the issues of Healy and his views on SSRIs.  It later transpired that his lawyer, Nina Gussack, who has represented Lilly on occasions, made it clear that Dr Nemeroff had been approached during the day previously when he was in Toronto and had made his views on Healy clear at that point and that he had been left with the impression that choices had been made there and then.  I was later to find out by email that there apparently had also been at least one phone-call to senior figures in Cornell in the days before I lectured there, making what appear to have been extraordinary claims and apparently suggesting the invitation be withdrawn. 

Following the lecture in Cornell, I arrived home from New York to find an e-mail waiting from David Goldbloom saying that:  “Essentially, we believe that it is not a good fit between you and the role as leader of an academic program in mood and anxiety disorders at the Centre.   While you are held in high regard as a scholar of the history of modern psychiatry, we do not feel your approach is compatible with the goals for development of the academic and clinical resource that we have. This view was solidified by your recent appearance at the Centre in the context of an academic lecture.”   

Many people reading this since have read a message that Goldbloom was concerned about pharmaceutical company monies to the University Department.  I was more struck the apparent implication that my analysis of the situation was correct (high esteem as a historian of modern psychiatry) but that a correct analysis was incompatible with employment in a modern biological psychiatry oriented department.  It appeared to smack of academic totalitarianism.  

Cheyenne

When it came to considering the ramifications of these events on my future career, there was a pressing issue to be addressed.  I was at this point involved in a case, Tobin versus SmithKline, which was due in court in Cheyenne, Wyoming, in May 2001.  I could envisage a situation where my first question on the witness stand would be about being sacked from the University of Toronto.  Where my first instincts would probably be in line with almost everybody else’s instincts in a situation like this, namely to lie low, it seemed I had little option but to do something.

On February 15th, I wrote to the Chair of the Ethics Committee at CAMH suggesting that in all probability neither CAMH, nor the university, nor I knew the full dimensions of what was happening.  For instance, I pointed out that in March 2000 there had been a special issue on Prozac brought out by the Hasting Center Reports, in which an article of mine called Good Science or Good Business made the points that we were treating more people than ever before, that Prozac could make people suicidal and that an increasing proportion of the therapeutics literature was ghost-written (Healy 2000).  Another even earlier article had argued that antidepressants and depression were almost unknown outside of mainstream psychiatry twenty years previously but we were now apparently in an Age of Depression (1999b).  The antidepressant market has in fact grown 800% by value in the 1990s and by 2000 had become a $10 billion market.  

It turned out that Lilly was one of the biggest private funders of the Hastings Center and following this article they withdrew their support.  The Hastings Center had my article re-reviewed.  One of these re-reviews said that essentially the only mistake I had made was in not going far enough in spelling out how much of the psychopharmacology literature was ghost written, how many of the clinical trials being run with psychotropic compounds ended up sealed and how much of the research was market oriented rather than designed to answer scientific questions.

CAMH had received a great deal of money from Lilly, SmithKline and other companies.  CBC television were later to report that in the previous year, around 50% of the Mood Disorders Program’s research money had come from pharmaceutical company sources.  Had this played a part in the decision?  Given that the points of concern in my lecture were similar to the points made in the Hastings Center article, there was a set of background issues against which the CAMH decision to breach my contract was likely to be judged in the wider academic domain.  Surely this provided grounds to talk before the situation got out of hand. 

My hope was that if they were better apprised of the dimensions of the problem, key figures in both CAMH and the university would be interested to meet up and try and find some solution.  I was due to speak in April at a meeting in Toronto and we could have met then.  But far from being interested to meet, the Chair of the Ethics Committee didn’t answer.  My February letter was answered by the Chair of the Board of Trustees in March - dismissively. 

I turned to the Canadian Association for University Teachers who wrote to the President of the University of Toronto expressing concerns.  The response from the President to CAUT was dismissive.  In April the issue began to run in the media in both Canada and the United Kingdom. 

The Tobin case began at the end of May.  Donald Schell was a man with a history of several relatively brief episodes of depression.  He had a prior history of an adverse response to Prozac in 1990.  He had then subsequently been put on Paxil by another physician in 1998 and 48 hours later had murdered his wife, Rita, along with his daughter and granddaughter who were staying with Don and Rita Schell before killing himself.  His surviving son-in-law, Tim Tobin, took out a case for wrongful death against Glaxo-SmithKline.

As part of my background research for this case, I had been given access to SmithKline’s paroxetine healthy volunteer archive.  It was clear from this that paroxetine caused agitation in around 25% of takers, that it did so in a dose-dependent way and on a challenge-rechallenge basis, and that there had been a suicide in the program.  It also caused physical dependence in one study in around 85% of subjects (Healy 2001b). 

This evidence of dependence was interesting in the light of SmithKline’s license to claim their drug was useful for the prophylaxis of depression.  This license was based on studies, which re-randomized patients to placebo, and interpreted the subsequent problems of those re-randomized to placebo as new illness episodes (Montgomery and Dunbar 1993).   

In the Tobin case, it became clear that studies had been terminated early with their results left unpublished.  It also became clear that despite a backdrop that gave serious grounds for concern, there had been a failure to test paroxetine or other SSRIs for the induction of possible suicidality.  Finally, the case raised questions about how much of a company’s defense in these SSRI cases depended on ghost-written, or company only authored publications, or how often when there was medical testimony it was based on tabulated figures provided to an expert rather than the raw data. 

In 2000 an article by Khan and colleagues in Archives of General Psychiatry had presented the following data from trials submitted to the FDA as part of the license applications for a series of recently released antidepressants:

Table 1. Incidence of Suicides and Suicide Attempts in Worldwide Phase 1 -3 Investigational Antidepressant Clinical Trials

 

Investigational Drug (Patient No.)

Patient

Exposure

Years

 

Suicide

No.

Suicide

Attempt

No

Sertraline hydrochloride (2053)

Active comparator (595)

Placebo (786)

 508

   91

 209

 2

 0

 0

  9

  1

  5

Paroxetine hydrochloride (2963)

Active comparator (1151)

Placebo (554)

1008

  218

   72

 5

 3

 2

 40

 12

   6

Nefazodone hydrochloride (3496)

Active comparator (958)

Placebo (875)

1018

  225

  204

 9

 0

 0

 12

   6

   1

Mirtazapine (2425)

Active comparator (977)

Placebo (494)

 672

 195

   71

 8

 2

 0

 29

  5

  3

Bupropion hydrochloride (1942)

Placebo (370)

----

----

 3

 0

----

----

All Investigational Drugs (12879)

Active comparator (3681)

Placebo (3079)

3206

  729

  556

27

 5

 2

 90

 25

 15

TOTAL (19639)

4491

34

130

But FDA documents (Brecher 1991; Lee 1991) revealed that some suicidal acts categorized as occurring on placebo in this table had in fact occurred during the washout period of clinical trials – the initial phase of trials when prior medication is halted before the patient is randomized to either placebo or active medication.  These gives rise to a revised table:

Table 2:  Incidence of Suicides and Suicide Attempts in Worldwide Phase 1 -3 Investigational Antidepressant Clinical Trials

Investigational Drug,

Patient No

Suicide No

Suicide
Attempt No

Sertraline hydrochloride

Active comparator

Placebo

Placebo Washout

 2053

   595

   786

 2

 0

 0

 0

  7

  1

  2

  3

Paroxetine hydrochloride

Active comparator

Placebo

Placebo Washout

 2963

 1151

   554

 5

 3

 0

 2

 40

 12

   3

   2

Nefazodone hydrochloride

Active comparator

Placebo

 3496

   958

   875

 9

 0

 0

 12

   6

   1

Mirtazapine

Active comparator

Placebo

 2425

   977

   494

 8

 2

 0

 29

  5

  3

Bupropion hydrochloride

Placebo

 1942

   370

 3

 0

----

----

All Investigational Drugs Above

Active comparator

Placebo

Placebo Washout

12,879

  3,681

  3.079

27

 5

 0

 2

 90

 24

  9

  5

 

Analyzing these suicidal acts by the traditional method, in terms of absolute numbers of patients, indicates a statistically significant excess of suicidal acts on new antidepressants as a group compared to placebo and a specific excess on paroxetine compared to placebo.  It also indicates that changing or withdrawing from medication can be more risky than is commonly noted.   

The data also suggest something else.  In the Tobin case, the defense appealed to the same data on paroxetine that appears in Table 1, which had previously appeared in two articles (Lopez-Ibor 1992; Montgomery et al 1995).  The failure to distinguish between placebo and washout suicidal acts suggests that the eminent authors of these articles may not have actually seen the raw data, which raises profound questions about the status of the psychopharmacology literature. 

On June 6th 2001, the Court came to a verdict, which found against Glaxo SmithKline.   In September 2001, I filed an legal action for breach of academic freedom against the University of Toronto and the Centre for Addiction and Mental Health. 

References

Brecher, M. 1991.  FDA Review & Evaluation of Clinical Data Original NDA 20-021, Paroxetine Safety Review June 19th.

Creaney, W., I. Murray, and D. Healy. 1991.  Antidepressant induced suicidal ideation.  Human Psychopharmacology 6: 329‑332

Healy, D. 1994.  The fluoxetine and suicide controversy. CNS Drugs 1: 223‑231

Healy, D. 1997.  The Antidepressant Era.  Cambridge Mass: Harvard University Press.

Healy, D., and M. Savage. 1998.  Reserpine Exhumed.  British Journal of Psychiatry 172: 376-378.

Healy, D. 1999a. Guest Editorial: A Failure to Warn.  Int J Risk & Safety in Medicine 12: 151-156.

Healy, D. 1999b. The Three Faces of the Antidepressants.  Critical Comments on the Clinical-Economic Framework of Diagnosis.  Journal of Nervous & Mental Disease 187: 174-180

Healy, D., C. Langmaak C., and M. Savage. 1999.  Suicide in the Course of the Treatment of Depression.  Journal of Psychopharmacology 13: 94-99.

Healy, D. 2000a.  Antidepressant induced suicidality.  Primary Care Psychiatry 6: 23-28.

Healy, D. 2000b.  Good Science or Good Business?  Hastings Center Report 30: 19-22.

Healy. D.,  M. Savage, P. Michael, M. Harris, D. Hirst, M. Carter, D. Cattell, T. McMonagle, N. Sohler, and E. Susser. 2001. Psychiatric bed utilisation: 1896 and 1996 compared.  Psychological Medicine 31: 779-790.

Healy, D. 2001c.  Trial testimony in Tobin vs SmithKline, Cheyenne, Wyoming, May 22nd 2001.

Healy, D. 2002.   The Creation of Psychopharmacology.  Cambridge Mass: Harvard University Press.

Khan, A., H.A. Warner, and W.A. Brown, 2000.  Symptom reduction and suicide risk in patients treated with placebo in antidepressant clinical trials.  Archives of General Psychiatry 57: 311-317.

Lee, H. 1991.  Statistical reviews on Sertraline for FDA, August 14th 1990 and January 31st 1991. 

Lopez-Ibor JJ. (1993). Reduced suicidality on paroxetine.  European Psychiatry 1 (Suppl 8): 17s-19s.

Montgomery SA, Dunner DL, Dunbar G (1995).  Reduction of suicidal thoughts with paroxetine in comparison to reference antidepressants and placebo.  European Neuropsychopharmacology, 5: 5-13.

Montgomery, S.A., and G.C. Dunbar. 1993.  Paroxetine is better than placebo in relapse prevention and the prophylaxis of recurrent depression.  International Clinical Psychopharmacology.  8: 189-195;

Tranter, R., H. Healy, D. Cattell, and D. Healy. (2002).  Functional effects of agents differentially selective to noradrenergic or serotonergic systems.  Psychological Medicine (in press). 


David Healy Qualified from University College Dublin in 1979.  Post-doctoral degree on Biochemical Markers in Depression.  Clinical Research Associate in psychiatry in University of Cambridge 1986 – 1990.  Reader in Psychological Medicine in University of Wales College of Medicine, Director of the North Wales Department of Psychological Medicine from 1992.  Former Secretary of the British Association for Psychopharmacology.  Author of over 120 peer reviewed articles and 12 books, including the reference history of the antidepressants – The Antidepressant Era, Harvard University Press – and The Creation of Psychopharmacology, Harvard University Press.  Other books include a three volume series of interviews – The Psychopharmacologists Volumes 1-3.  

Correspondence:

North Wales Dept of Psychological Medicine
University of Wales College of Medicine 
Hergest Unit
Bangor 
North Wales, U.K. LL57 2PW

Email: Healy_Hergest@compuserve.com