Court Filing Makes Public My Previously Suppressed Analysis of Paxil’s Effects

[Guest guest]

SPECIAL REPORT

Court Filing Makes Public My Previously Suppressed Analysis of Paxil's

Effects

Peter R. Breggin, MD

 

 

Ithaca, NY

In late 1999 I was asked by attorney Don Farber to be the medical

expert in a product liability case brought by the family of Reynaldo

Lacuzong against the drug company Glaxo SmithKline (GSK) in California.

 

Mr. Lacuzong was a machine operator with no prior history of serious

mental illness, violence, or suicidality before he was prescribed a

relatively small dose of 10 mg of Paxil (paroxetine).

 

Almost immediately after starting the Prozac-like selective serotonin

reuptake inhibitor (SSRI) antidepressant, he developed akathisia—an

inner agitation accompanied by a compulsive hyperactivity—as well as

other manic-like signs of irritability and anxiety.

 

Antidepressant-induced akathisia is known to be associated with

violence, suicide, psychosis, and an overall mental deterioration

(American Psychiatric Association, 2000, pp. 800–802). Depression with

drug-induced agitation can produce similar results.

 

On the third day of taking Paxil, Mr. Lacuzong drowned himself

and his two small children in a bathtub.

As a medical expert, I was empowered by the court to examine hundreds

of cartons of drug company files contained in GSK's sealed record room.

 

These files included Food and Drug Administration (FDA) correspondence

and all of the company's worldwide clinical trials and adverse drug

reports for Paxil.

 

On July 21, 2001, my report in the form of an affidavit was sent to

the judicial arbitrator in the case. It addressed GSK's practices in

the development and marketing of Paxil, and in particular its alleged

withholding or manipulation of information about the drug's dangerousness.

 

Based on GSK's proprietary files that have to this day never been made

public, my report examined many factors, including (a) how quickly

after the first dose can Paxil cause severe adverse reactions; (b) the

actual rates of akathisia; © the actual risk of overstimulation

causing agitation, irritability, and manic-like symptoms; (d) the

actual rates of suicidality in adults; and (e) promotional claims made

for the drug.

 

The case against GSK was eventually " resolved " to the satisfaction of

GSK and the Lacuzong family. GSK denied and continues to deny all of

the allegations of negligence in developing and marketing Paxil.

 

My impression is that a substantial amount of money was involved in

the resolution of the case, although the amount was not disclosed. GSK

at that time refused to unseal its records or to allow me to make

public my findings, regardless of their significance for the FDA,

medical profession, and public health.

 

On June 23, 2005, my report in the Lacuzong case was filed as a part

of a motion in another Paxil case, Moffett v. Glaxo SmithKline, in the

United States District Court for the Southern District of Mississippi.

Because it was filed in the public record, my report is now

Ethical Human Psychology and Psychiatry, Volume 8, Number 1, Spring

2006 © 2006 Springer Publishing Company 77 available to the public,

and I am able to comment on the details.

 

The complete version appears on my website www.breggin.com. The report

should prove useful to the FDA, health practitioners, scientists,

researchers, attorneys, consumers, and anyone concerned about

how drug companies function in the development and marketing of their

products.

 

In the meanwhile, the FDA has recently acknowledged many of my

original observations about the stimulating effects of all of the

SSRIs like Paxil and Prozac, as well as other new antidepressants such

as Effexor. I first warned about these effects in Toxic Psychiatry

(Breggin, 1991) and then in subsequent peer-reviewed articles and

books (see for example, Breggin, 1997, 2001, 2003).

 

As of 2005, the FDA now requires the drug manufacturers to place

elaborate warnings on their labels concerning the potential of these

drugs to cause stimulating effects, including agitation, anxiety,

irritability, emotional lability, aggression, hostility, and mania.

 

The labels must also include a warning about increased suicidality in

children. Without coming to a scientific conclusion, the FDA has also

warned about and begun to investigate the problem of antidepressant

suicidality in adults.

 

Among other things, my report in Lacuzong deals in detail with

antidepressant-induced suicidality in adults and how the drug company

handled that data.

 

The following excerpts include Sections XI–XIV of the report that can

be found in its entirety on www.breggin.com. These excerpted sections

focus on Paxil-induced suicidality in adults—the subject of a current

FDA investigation. Other sections deal with numerous issues including

FDA correspondence criticizing the company, advertising and promotion,

the drug label, rates of adverse psychiatric effects after only a few

doses, akathisia, and the stimulant or activation continuum. A

lengthy, detailed summary and conclusion can be found in the complete

report.

 

EXCERPTS FROM THE REPORT AND AFFIDAVIT OF PETER R.

BREGGIN, MD, IN THE CASE OF LACUZONG V. GLAXO

SMITHKLINE

XI. Evaluating Errors in the Compilation of Suicide Data

1. Suicide Attempts: U.S. Clinical Trials. A total of 14 suicide

attempts were reported

in the U.S. clinical trials. None were completed suicides. An overview

is presented in Table

1 (PAR Safety Summary 20-Nov-1989, p. 203, stamped p. 297).

 

Note that the rate for suicide attempts on paroxetine approaches 1%,

which the FDA

considers " frequent. " Also note that the rate for suicide attempts on

paroxetine is 3.8 times

higher than for placebo and 3.6 times higher than for the comparison

antidepressants (tricyclics).

 

Furthermore, the suicide attempt on imipramine is listed as a

" possible suicide "

(p. 211, stamped 306).

 

In regard to the onset of suicide attempts, one patient (117A-004, p.

200, stamped 291)cut himself on the third day of Paxil: " One day 3

this patient attempted to slash his wrists and abdomen and was

withdrawn from the study. " Also note that case 647 002 (above)

made attempts on days 1, 8, and 15.

 

This all-important United States data is not presented in the text of

SKB's (now Glaxo SmithKline or GSK)April 29, 1991, report for the FDA,

" Suicidal Ideation and Behavior:

Analysis of the Paroxetine Worldwide Clinical Database. " To hide the

U.S. data within worldwide data was extremely misleading.

 

78 Special Report

2. Leaving Out Two Non-U.S. Suicide Attempts. There is evidence that

some suicide attempts were omitted from the calculations sent to the

FDA. In the report " Adverse experiences which occurred during active

treatment. Non-US Phase II-III studies "

 

(Appendix V.1), I located two patients that appear to have been left

out of the summaries of non-U.S. suicide attempts. Case 647 002

(Volume 420, p. 157) made three suicide attempts on days

1, 8, and finally on day 15, when the drug was stopped. The first two

were considered " related " and the third " possibly related. " Also, case

1 113 120 (Volume 420, p. 157) was considered " definitely drug related. "

 

These two attempted suicides do not appear in the complete list of 40

in the April 29, 1991, suicide report (pp. 17–18), or in any other

source that we have located. This brings the total of non-U.S. suicide

attempts to 32.

3. Leaving Out Two Non-U.S. Completed Paxil Suicides. Two non-U.S.

completed suicides appear to have been left out of all official

reports, including the April 29, 1991, suicide report. The missing two

are found in Appendix 5.4.2—Summary of Deaths Occurring in

Paroxetine-Treated Patients (unnumbered page, SB 0000044).

See Table 2 for the seven cases with their complete descriptions under

the heading of " Cause of Death and Comments. "

Special Report 79

TABLE 1. Overview of Attempted Suicide: U.S. Dataa Paroxetine Placebo

Other ADb

N = 1,562 N = 497 N = 464

Drug overdose 9 0 1

(imipramine)

Defenestration 2 0 0

Self-inflicted injury 1 0 0

Suffocation 0 1 0

Totals 12 (0.77%) 1 (0.20%) 1 (0.21%)

aTable numeration added. bAD = antidepressant (imipramine).

TABLE 2. Seven Suicides Listed Under " Cause of Death and Comments " a

Case Number Duration, days Cause of Death and Comments

DFG124/12* ? Suicide: Method—Overdose with doxepin

HDUK13/26* 144 Suicide: Method—Hanging

29060/149* 18 Suicide: Method—Overdose . . . 6 days

after discontinuing paroxetine

HP-82-47/3* 47 Suicide: Method—Drowning

058/022** ? Suicide: Method—Unknown

083.003.1090** 8 Suicide: Method—Hanging

2206.005/605* 150 Suicide: Hanging

aTable numeration and heading added.

 

Since only five non-U.S. suicides are listed in any of the tables or

reports, it is apparent that there are two missing. Two more should be

added to the suicide completed counts (see below). I cross-checked

these numbers and have found that five are included (see single asterisk

[*]) in the official lists of suicides. We need to obtain the two

missing cases (see double asterisk [**]).

Appendix 5.4.2 appears to be based on the Summary Basis of Approval

data (SBA), which draws from the NDA. It was part of a list of 15

deaths described in an August 25, 1992, memo entitled " Miscellaneous

Requests " from Thomas P. Laughren, MD, of the FDA to Thomas Donnelly,

Jr., PhD, of SKB. He notes he is adding one, 083.003.1090,

from the safety update, which is also a part of the original NDA.

We need to inquire about any further correspondence or corrections

concerning this list.

4. Adding Two Placebo Run-In Completed Suicides to the Non-U.S.

Studies. In the suicide report the following two suicide cases are

listed: 7119.062 and 7119.009.

However, both of these occurred during the placebo run-in (also called

placebo wash-out) phase. The cases can be found summarized in the PAR

Safety Summary 20-Nov-1989 (7119.062 on p. 202c, stamped p. 296, SB

0000544; 7119.009 on p. 202b, stamped p. 295, SB 0000543).

There is no question that placebo run-in is a euphemism for placebo

wash-out. In the April 29, 1991, suicide report a footnote states,

" Suicides were committed during the placebo wash-out phase of an

active control study. These two acts were committed 2 days and 7

days prior to the baseline evaluation, i.e., -2 and -7 days. "

Adverse drug effects are never reported from the placebo wash-out phase.

 

Indeed, suicide and suicide attempts are probably the only supposed

adverse drug effects reported from the placebo wash-out. The placebo

wash-out period is not a part of the controlled clinical trials.

 

It occurs before the randomization. All patients are lumped into them.

Furthermore, many of the patients are very likely suffering from

withdrawal from other drugs they were previously

taking for depression.

 

The inclusion of these suicides into the placebo comparison group was

misleading to the extreme. They must be removed from calculations

pertaining to a comparison between suicides on Paxil and on placebo.

5. Including Two Placebo Run-In Suicide Attempts in Non-U.S. Studies.

The worldwide data for suicide attempts also include placebo run-in

data. This is confirmed in Table XI.21, Attempted Suicides and

Overdoses—Worldwide Data (PAR Safety Summary 10-

Nov-1989, p. 206, stamped page 300, SB 0000548).

 

Exactly as in the case of including completed suicides from the

placebo wash-out phase, the inclusion of two placebo run-in

patients in the non-U.S. suicide attempt category is misleading and

fraudulent. The two placebo run-in patients must be excluded from the

non-U.S. and worldwide data.

 

XII. Re-Analysis of the Suicide Data

1. Re-Analyzing Non-U.S. Completed Suicides. Various SKB documents,

including the April 29, 1991, suicide report, only list five completed

suicides. As described above, we have found an additional two for a

total of seven. Therefore the completed suicide rate for Paxil is

seven in a population of 1,401 patients for a rate of 0.499%.

 

As also described above, we found that two placebo wash-out completed

suicides were wrongly counted in the suicide rate for placebo. The

true occurrence for completed suicides

80 Special Report in the placebo group is 1 in 544 for a rate of

0.180%. The suicide rate on Paxil is therefore 2.7 times that on placebo.

2. Creating a New Category of Suicidal Behavior or Suicides, Attempted

and Completed.

The 5 completed Paxil suicides (acknowledged by SKB) must be added

together with the 42 (from Table XI.21) attempted suicides to create

the category of Suicidal Behavior or Suicides, Attempted and Completed.

 

The category contains, at the least, 47 cases of

suicidal behavior (42 + 5 = 47). SKB's analysis obscures and hides the

actual rate of suicidal behavior by evaluating attempted and completed

suicides as separate entities.

 

We also need to know the overall rate of suicidal behavior.

Based on this analysis, the rate of suicidal behavior is 47 out of

2,963 for a rate of 1.58%.

If we add the additional two completed suicides that seem to have been

left out of the data, we now have 49 (47 + 2 = 49) suicidal behaviors

out of 2,963 for a rate of 1.65%. Whether we use the 1.58% figure or

the 1.65% figure, this combined category of suicidal behavior

is far more meaningful than the split categories of suicide attempts

and suicides completed.

 

It was grossly misleading not to create a combined category.

The above calculations were based on the assumption that there were 42

suicide attempts as indicated in the original NDA. If we added the two

suicide attempts that appear to have been left out of the data, there

are at least 44 total suicide attempts. The corrected

total for combined suicidal behavior on Paxil then becomes 51 (44

suicide attempts + 7 suicides = 51). Fifty-one out of 2,963 produces a

rate of 1.72% for suicidal behavior on Paxil.

 

3. Re-Analysis of the Worldwide Comparisons for Suicide Attempts. We

have already found that two attempted suicides on Paxil were

apparently not included in the worldwide

calculations. As described above, this raises the original NDA figure

from 42 to 44 for attempted suicides out of 2,963 cases, for a rate of

1.48%. In addition to undercounting suicide attempts on Paxil, SKB

overcounted placebo-related suicide attempts.

 

For placebo, three suicide attempts are listed. But as we have

documented, the correct number for placebo suicides is only one for

the worldwide group. The other two suicide attempts

were placebo wash-out cases. That makes the placebo suicide attempt

rate a mere 1 out of 554 for a rate of 0.18%.

 

Thus the corrected comparison indicates a 1.48% rate of suicide

attempts on Paxil compared to a 0.18% rate of suicide attempts on

placebo worldwide. Thus suicide on Paxil was 8.2 times higher than the

rate of placebo.

 

4. Hiding the Frequency of Suicide Worldwide in the April 29, 1991,

Suicide Report.

In the " Discussion and Conclusions " of the April 29, 1991, report (SB

0000819, report pp. 12–13) states the following conclusion:

The incidence of attempted suicides did not differ substantively among

the three treatment groups (paroxetine, placebo, active controls).

However, the report never deals with the U.S. clinical trials as a

separate entity. They show a significantly higher suicide attempt rate

for Paxil than for the other antidepressants

or placebo.

 

Furthermore, there is no overall category of Suicidal Behavior or

Suicides, Attempted and Completed. Therefore, when counting suicide

attempts, suicides completed are Special Report 81

excluded, badly misrepresenting the data. In addition, there appear to

be two unreported suicide attempts and six unreported completed

suicides worldwide.

 

Finally, as already noted, the worldwide figure is distorted by

miscounts in both the Paxil and placebo categories.

The April 29, 1991, suicide report also contains different numbers

from the NDA.

 

We find that the total number of paroxetine suicide attempts has been

inexplicably reduced from 42 in the NDA to 40 two years later, while

the total number of placebo suicide attempts has been inexplicably

increased from 3 to 6. These manipulations of course favor

the interest of the drug company. The April 29, 1991, report in fact

states that it has based itself on the original NDA data, that is,

" using data which were submitted at the time of the New Drug

Application for paroxetine " (p. 1, SB 0000003). But the NDA data

differ to the disadvantage of SKB.

 

XIII. Follow-Up of U.S. Suicide Attempt Cases

I was able to track many but not all of the individual case numbers

listed in the compilation of suicide attempts (Table XI.19 from PAR

Safety Summary 20-Nov-1989, p. 203, stamped p. 297). The cases were

found separately in a book-length document, " Narrative

of US Patients with Potentially Clinically Significant Events "

(Appendix I.1 of NDA 20031, 409, November 1989).

 

They indicate that the suicide attempts often occur in a context of

various other distressing adverse drug reactions but sometimes occur

without any other serious adverse effect. This contrasts with the

non-U.S. data on completed suicides which indicate that the five we

could track were all related to central nervous system

adverse drug reactions, including akathisia and stimulation.

(1) 02-04-089 (p. 37). This patient had been taking Paxil 20 mg for 40

days. " Adverse clinical experiences . . . were moderate dizziness and

lack of energy (probably drug related), and moderate headaches

(possibly drug related). "

(2) 04-01-009 (p. 192; SB 0000571). This patient elected to switch

from a tricyclic to Paxil. After 193 days the patient was taking 50 mg

and experienced the following adverse

reactions: " clenching of teeth, " dry mouth, decreased libido,

inability to achieve orgasm, nausea, diarrhea, urinary retention,

" weakness in legs, " " twitching of left cheek, " lightheadedness,

anxiety, " speedy feeling, " dizziness, " tingling, " lethargy, headache,

and decreased concentration.

 

(3) 04-02-056 (Volume 409, p. 260). This patient was taking Paxil 40

mg and at 19–20 days made self-inflicted scratches. The patient was

given ECT [so probably experienced a worsening of depression]. Other

than dry mouth, no other ADRs were reported.

(4) 04-06-96. This patient was on 30 mg of Paxil for 116 days. The

patient could not be located in the " Narrative of US Patients with

Potentially Clinically Significant Events. "

(5) 05-01A-030 (Volume 410, p. 65). This 23 year old patient was

taking Paxil 50 mg and attempted suicide twice. The two attempts were

counted only once. " The patient required hospitalization because of

excessive ethanol use with violent and unpredictable

behavior. " She intentionally overdosed.

(6) 05-01A-075. This patient was a 37 year old female taking Paxil 40

mg for more than three years. She was not located in the " Narrative of

US Patients with Potentially Clinically Significant Events. "

(7) 05-02B-019 (Volume 410, p. 124). This patient was taking Paxil 50

mg for 57 days when the overdose occurred. " Adverse experiences

reported during the study were 82 Special Report

mild rash, diarrhea, `shakiness' (possibly drug related), and an

overdose. " She took 20–50 unknown pills and was hospitalized.

(8) 05-02F-002 (Volume 410, p. 151). This patient was taking Paxil 40

mg for 38 days and attempted suicide. No other ADRs were reported.

(9) 07-01A-001. This person was taking Paxil 40 mg for 20 days. The

case could not be located in the " Narrative of US Patients with

Potentially Clinically Significant Events. "

(10) 09-01A-005 (Volume 410, p. 196). This patient was taking Paxil 40

mg and overdosed at 7 days. She was experiencing " moderate drowsiness,

tremulousness, severe nausea (probably drug related), and overdose. "

She overdosed for a second time 7–8 days later.

There were therefore two overdoses, one during drug exposure, and one

apparently within a week after withdrawal.

(11) 09-01E-260. This patient was taking Paxil 10 mg for 60 days. The

patient could not be located in the " Narrative of US Patients with

Potentially Clinically Significant Events. "

(12) 09-01J-573 (Volume 410, p. 279). This patient was taking Paxil 10

mg according to the summary (p. 298) but taking 20 mg according to

this case report. The drug exposure was listed as 26 days but appears

to have been 30 days. The patient " jumped from second story window "

and " received multiple fractures. "

In addition to these 12 Paxil patients who attempted suicide (for a

total of 14 attempts), there was one attempted suicide on imipramine

and one on placebo. They follow:

(13) 04-06-088 (Volume 410, p. 50). This patient was taking imipramine

225 mg for 61 days. The patient was listed as a " possible suicide

attempt. " " He reportedly had taken an unknown quantity of `pills' and

was intoxicated. " In fact, this is probably not a suicide

attempt.

If this case is discarded, there are no other cases of suicide attempt

on the comparison drug and the ratio becomes 12–14 to 0. It appears

that the drug company attempted to cover up the higher

rate of suicide attempts on Paxil by including this unlikely case of a

suicide attempt.

(14) 02-01-009 (volume 410, p. 5). This patient was on placebo for 6

days. The case is described as " a suicide gesture by suffocation. Her

husband prevented her suicide. " Notice that this case is a " gesture. "

I found no " gestures " included in the Paxil group.

If this case is discarded, as well as the one imipramine case, then

there were 12–14 suicide attempts among 12 patients on Paxil and none

on placebo or on imipramine.

XIV. Increasing Evidence of Suicidality on Paxil

On 1.14.00 the FDA wrote a 3-page letter to Thomas Kline of SKB

suggesting a label change. The FDA recommends a label change under

" Overdosage/Human Experience. "

Since the introduction to the US, 342 spontaneous cases of deliberate

or accidental overdose with paroxetine have been reported worldwide

(circa 1999). Seventeen involved Paxil by itself. There were 48

fatalities.

This issue is even more serious than the FDA indicates, since there

are obviously a large number of suicide attempts in this group.

REFERENCES

American Psychiatric Association. (2000). Diagnostic and statistical

manual of mental disorders (4th

ed., text revision) (DSM-IV-TR). Washington, DC: Author.

Special Report 83

Breggin, P. (1991). Toxic psychiatry. New York: St. Martin's Press.

Breggin, P. (1997). Brain-disabling treatments in psychiatry. New

York: Springer Publishing Company.

Breggin, P. (2001). The antidepressant fact book. Cambridge, MA:

Perseus Books.

Breggin, P. R. (2003). Suicidality, violence, and mania caused by

selective serotonin reuptake inhibitors:

A review and analysis. Ethical Human Sciences and Services, 5, 225–246.

84 Special Report