04_EHPP_8(1)_leo.qxd 1/27/06 3:08 PM Page 29
Ethical Human Psychology and Psychiatry, Volume 8, Number 1, Spring 2006
The SSRI Trials in Children:
Disturbing Implications for
Jonathan Leo, PhD
The recent announcement by the Food and Drug Administration (FDA) requiring phar-maceutical companies to warn patients about the increased likelihood of suicidal thoughtswhen taking antidepressants was largely due to the recent availability of data that had goneunreported in the original research reports. The current article is a summary of the com-parison between the published literature and the recently released data available on the FDAweb site, with a focus on Prozac, Paxil, and Zoloft. The discrepancies between the two ver-sions suggest that the scientific community was not given enough information in the pub-lished medical literature to make adequately informed decisions about the optimal methodfor treating emotionally distressed children. There are many voices that want to blame theFDA for its role in the widespread use of these medications. The current article focuses onthe role that academic medicine, played in the process whereby these medications becameso widely accepted.
SSRI; suicide; pediatric depression; FDA
“Confusion, manipulation, and institutional failure” was the recent summary on
’s editorial page about the research into selective serotonin reup-take inhibitor (SSRI) use in childhood depression (Editors, 2004). The edi-
tors’ conclusions were based on revelations that pharmaceutical companies had selectivelyreported favorable research about the use of antidepressants in children. However, the med-ical community makes a mistake if it believes that the current problem is just one of un-published data, which the pharmaceutical companies have kept hidden. If it were only asstraightforward as a relatively few acts of corporate irresponsibility, then the solution wouldbe fairly simple, but virtually everyone knows it is more complicated. The SSRIs are notthe only drugs involved and, moreover, the pharmaceutical companies are not the onlyplayers involved. The more important problem is that for 10 years the system of academicmedicine failed to the point that the academic journals were circulating a myth about thebenefits of psychotropic drugs for children that had little to do with the truth.
2006 Springer Publishing Company
04_EHPP_8(1)_leo.qxd 1/27/06 3:08 PM Page 30
In October 2004, the Food and Drug Administration (FDA) announced that commonlyused antidepressants likely lead to an increase in suicidal thoughts or actions in some chil-dren and teens; they also announced that the same problem might exist for adults. Almost1 year prior to this, the Medicines and Healthcare products Regulatory Agency (MHRA),the British equivalent of the FDA, effectively banned the use of these drugs, except forProzac, in children and adolescents under 18 years of age. Ten years ago, the market for an-tidepressants in children was practically nonexistent, 5 years ago it was booming, and now,it seems to be in jeopardy. Considering how much money we have spent over the pastdecade on the search for the causes and treatment of mental anguish, who would havethought we would end up with a warning label saying that these drugs might lead to an in-creased rate of suicide?
In the late 1980s, adults started to use the selective serotonin reuptake inhibitors
(SSRIs) for depression, and since then the brand names, such as Prozac, Paxil, and Zoloft,have become household names. Indeed, three of the seven most commonly used drugs arenow mood elevators. By the early 1990s, it didn’t matter that they were not officiallyapproved for use in children: they were commonly given to children as young as 6 yearsold. In some cases prescriptions were written for infants under twelve months (Grinfeld,1998).
Although you would never know it from a perusal of the mainstream psychiatry journals
over the last decade, when the SSRIs first came on the market in the late 1980s there werehints of increased suicidality in people taking the drugs (Healy, 2003a). It is not surprisingthat some psychologically distressed people commit suicide; after all, a distressed state is apredisposing factor for suicide. So, naturally, to accusations that the drugs cause some peo-ple to take their own lives, the companies’ response has always been: It’s the disease, notthe drug. While it may seem hard to disentangle the role of these two variables, the diseaseand the drug, there is a way, and it has been done. The place to begin is with the compa-nies’ own efficacy studies.
These studies started with a group of depressed patients, which was divided in half. Half
the group was given a drug, and the other half a placebo. According to the FDA data, therewas an increase in suicidality in the drug-exposed group. Furthermore, this is not an iso-lated finding, but is the case for almost every single antidepressant studied—in adults, noless (Healy, 2003b). No longer can we simply blame the disease: the drugs appear to beplaying a role in making some people more likely to take their lives.
In the late 1990s, the first major scientific papers investigating the use of these drugs in
children, funded by the drug companies, started to appear in the medical literature. Thestudies made a case for these drugs as being safe and effective for the treatment of child-hood distress. The studies subsequently became part of the companies’ submissions to theFDA to get the drugs approved for children.
However, drug companies do not just supply the FDA with references to published
papers, they supply significantly more detail than what is available in the scientific pa-pers. The version submitted to the FDA is not available immediately to the public, but,eventually, some of the data finds its way onto the FDA web site, or becomes availabledue to court proceedings. Only when all the data is in the public realm can a comparison
04_EHPP_8(1)_leo.qxd 1/27/06 3:08 PM Page 31
between the two versions be made: the one published in the medical literature and theunpublished account submitted to the FDA. It was this embarrassing mismatch betweenthe published and unpublished data that forced the FDA to step back into the fray.
To win approval of a drug, for a given condition, in a certain population, a drug com-
pany must submit two positive studies to the FDA. To obtain its two positive studies, thecompany might do five, six, or as many as ten studies. Even if the majority of the studiesare negative, as long as two are positive the drug can be awarded FDA approval. For anygiven study, the basic idea is that the benefits should outweigh the risks. It is taken forgranted that every drug will cause some sort of side effects, but if they are minimal andthe benefits are significant, then the FDA, the medical community, and the patients allagree to live with the risks. Out of seven requests to the FDA for approval of a specific anti-depressant to treat childhood depression (Prozac, Paxil, Zoloft, Celexa, Remeron, Effexor,Serzone), Prozac was the only drug that was approved. While not
approved for the treat-ment of depression, Zoloft was approved for the treatment of Obsessive Compulsive Dis-order (OCD) in children.
In terms of benefits, the researchers need to show that patients treated with the drug do
significantly better than patients treated with a placebo. If the FDA decides that the ben-efits outweigh the risks then the drug is officially approved. Importantly, once a drug is ap-proved, for, perhaps, depression in adults, there is nothing stopping doctors from prescribingthe drug to other groups of people—to children, for instance. Nor is there anything to keepdoctors from prescribing the drug for other psychiatric conditions, such as, OppositionalDefiant Disorder (ODD). Although Prozac was not approved for children until 2003, itwas perfectly legal for doctors to prescribe it, off-label, to children prior to official FDA ap-proval. Likewise, even though the other SSRIs are still not approved for use in children, itis perfectly legal to use them.
In fact, the child psychiatry profession fully endorsed the use of these drugs well be-
fore the FDA approved them, and, in an even odder twist, the profession endorsed theuse of them well before any of the major studies in children were even published(Koplewicz, 1997). It appears that one reason for doing the studies in the first place wasto justify already well-accepted prescribing patterns. If a trend is created “because every-one else is doing it” then it appears that the child psychiatry profession’s use of thesedrugs in the late 1990s more closely resembled a trend instead of a logical scientificundertaking.
How did all this happen? Naturally, there are no easy answers, but, as a start, one needs
to look at the studies asserting that these drugs are safe and effective. In the following dis-cussion, the papers written for the scientific community are compared to the FDA docu-ments (Mosholder, 2004). For the most part, the field consists of about 12 papers, spanningalmost a decade, which, when stacked together, fit nicely into a three-ring binder. Thestudies involving the three most common selective serotonin reuptake inhibitors (SSRIs)—Prozac, Paxil, and Zoloft—serve as excellent examples of how, at every step of the way,the benefits were overestimated and the risks underestimated. According to Healy, “Thereis probably no other area of medicine in which the academic literature is so at odds withthe raw data” (Healy, 2004, p. 10). On one hand, debating the details from any one study,about how one or two children’s side effects should be categorized, might seem trivial. Onthe other hand, when the literature on the use of antidepressants in children is lookedat as a complete body of work there is a problematic track record, which is anything buttrivial.
04_EHPP_8(1)_leo.qxd 1/27/06 3:08 PM Page 32
STUDIES SUBMITTED FOR FDA EVALUATION
The lead author of the two studies submitted to the FDA as part of Eli Lilly’s successful re-quest to get Prozac approved for children was Graham Emslie, a professor at The Univer-sity of Texas Southwestern Medical Center at Dallas. The first study, funded by the NationalInstitute of Mental Health (NIMH) and published in 1997, reported that “side effects, asa reason for discontinuation, were minimal” (Emslie et al., 1997, p. 1033). There is no men-tion in the paper about any children attempting suicide. However, in the FDA’s “MedicalReview of Prozac,” written in 2001 but not made public until 2003, there is a discussionabout two children on Prozac attempting suicide (Center for Drug Evaluation and Research,2001; Leo, 2004). Thus, between 1997 and 2003, doctors reading the published paper andtrying to decide if Prozac should be used in children were not given all the information.
Even by 2003, only those doctors who regularly peruse the FDA web site would have knownabout the two suicide attempts in the group of 48 Prozac-treated patients.
Under government pressure, the drug companies have released their unpublished data,
while NIMH has still not explained the discrepancy between its own published and un-published versions of the first Prozac for children paper. Also unclear are the details of howthe data from this NIMH study became part of Eli Lilly’s package that was submitted to theFDA for the evaluation of whether or not Prozac should be used in children. There is noth-ing wrong with Lilly simply using data in the public realm, but the data Lilly submitted tothe FDA went beyond what was in the public realm, that is, the two attempted suicides.
Eli Lilly’s access to NIMH’s unpublished data suggests a close relationship between NIMHand the drug companies. The obvious question is: Did Lilly pay NIMH or the researchersfor data that was the product of a taxpayer-funded research project? Regardless of the mon-etary questions surrounding the first Prozac study, by the time the second Prozac study waspublished in 2002, Emslie and his coauthors were either Lilly employees or paid consul-tants. And now, as of 2004, Emslie is still receiving money from NIMH to continue in-vestigating these medicines in children.
When an investigator is simultaneously receiving funds from a nonprofit institution and
a for-profit company—one organization primarily committed to finding the truth, and oneorganization primarily committed to satisfying shareholders—it has to be extremely diffi-cult to keep these allegiances straight. When a professor in this situation, at a scientificsymposium for instance, claims that “Studies show . . .” is the professor referring to the pub-lished studies, or all the studies? As we are seeing, this qualifier has now become one of theessential pieces of information that goes into evaluating a scientific claim.
What is clear, though, is that, in general, all these studies are designed to give the drug
the best chance of coming out ahead. For instance, the two Prozac studies, and several of theothers, included a placebo washout phase, which involved putting all the patients on aplacebo for a specified time period and then dropping those patients that improved. In thefirst Prozac study, 10 children were dropped from the study during the placebo washout.
The second Prozac study also had a unique twist, which consisted of a run-in phase to
preselect for drug responders (Dubitsky, 2004; Emslie et al., 2002). All the Prozac-treatedchildren in this study were given 10 mg for the first week, and children who did not re-spond, or who had negative responses, could then be dropped from the study (p. 1206).
At the start of week 2, the dose was increased to 20 mg. The subsequent statistical analysisonly used children who had had at least one week of treatment with 20 mg (p. 1208). Thus,
04_EHPP_8(1)_leo.qxd 1/27/06 3:08 PM Page 33
before the study even started, there was a mechanism in place to maximize any differencebetween the drug and placebo groups— the placebo group was preselected for nonrespon-ders
, while the drug group was preselected for responders
. Yet even with this advantage, forthe prospectively defined primary outcome measure, 65% of the children on Prozac had abeneficial response compared to 53% of the placebo patients, a result that was not statis-tically significant. It was only by looking at other measures that clinical significance wasfound; on the patient- and parent-rated scales there was no advantage to Prozac, but on oneof the clinician-rated scales there was a slight advantage to Prozac. Although Russell Katzof the FDA wrote, “one could argue that this post hoc choice of primary outcome is inap-propriate,” the FDA accepted the post hoc change and approved Prozac for children inJanuary 2003 (Center for Drug Evaluation and Research, 2002, p. 13). It was the only an-tidepressant that the FDA ever approved for use in childhood depression.
In 2003, when the British MHRA banned the use of these medications in children and
teens, Prozac was spared. Not because the suicide profile was noticeably different, but be-cause Prozac appears to offer more benefit than its competitors (although even these pointsare open to debate). Yet, according to the data, the efficacy of Prozac is not very significant.
Using the researchers’ own criteria for defining an “improvement,” and taking into accountthe large placebo effect, for every ten children it is given to, it maybe helps one patient.
In 2001 The Journal of the American Academy of Child and Adolescent Psychiatry
published astudy that found that Paxil was appropriate for the treatment of emotional distress in chil-dren. The paper, authored by the leaders of the child psychiatry profession, concluded that,“Paroxetine is generally well tolerated and effective for major depression in adolescents”(Keller et al., 2001, p. 762). At first, this paper seemed to justify the prescribing patternsof the past several years.
But, at the time the paper was published, it was clear to some people that there were
major problems with the study (Jureidini & Tonkin, 2003). First, the study provided littleevidence that Paxil was better than placebo. For the primary endpoint, the difference be-tween the Paxil-treated and placebo groups was not significant (Laughren, 2004, p. 5).
Secondly, there were significant side effects in the Paxil-exposed children compared tothose in the placebo group.
While the authors of the study vigorously defended their paper against the critics, prob-
lems arose when internal, confidential company documents surfaced, supporting the crit-ics instead of the company’s own paid experts. The memo, written in October 1998 andstamped, “For Internal Use Only,” was a summary of the same Paxil study. It acknowledgedthat the study, at that point referred to as Protocol 329, had not demonstrated efficacy forPaxil and that “it would be commercially unacceptable to include a statement that effi-cacy had not been demonstrated, as this would undermine the profile of paroxetine”(GlaxoSmithKline, 1998).
In the published account of Protocol 329, the authors included an extensive list of 32
side effects such as nausea, vomiting, chest pain, and so forth. Although, ever since theSSRIs were approved for adults, there has been the suspicion that they lead to increasedthoughts of suicide (Healy, 2003b), there is no mention of any children having any suici-dal thoughts or attempting suicide in the published version of Protocol 329. Interestingly,six of the children on Paxil had “emotional lability.” In reality, some of these “emotionallylabile” children actually had suicidal thoughts (Healy 2004). At the urging of the British
04_EHPP_8(1)_leo.qxd 1/27/06 3:08 PM Page 34
government, GlaxoSmithKline acknowledged that in Protocol 329’s Paxil-treated group,there were 5 children out of 93 who had suicidal thoughts; in the placebo group none ofthe 89 children had suicidal thoughts. Prior to this revelation the average physician wouldhave been left completely ignorant of this vital statistic.
The investigative news show Prime Time Live
recently broadcast a segment about
GlaxoSmithKline distributing a memo (different than the one mentioned above) to itssales force touting Paxil’s, “remarkable efficacy and safety in the treatment of adolescentdepression” in 2001, at a time when the company was fully aware that there were problemswith Paxil (ABC, 2004). But far more problematic for the medical community than a salesforce being given a slanted view from sales managers is that doctors were given the sameslanted view in medical journals from scientists. If the company is at fault for concealingdata from its sales force, then what about the scientists who wrote the papers? Whether youbelieve it was the pharmaceutical salespeople or the medical school professors who con-vinced clinicians to use these medicines, probably says a lot about your views of medicinein general. Shouldn’t the media’s focus really be on the journals, not the marketingmemos?
It would be interesting to know the timeline of events, and the nature of the scientists’
involvement with the Paxil study. For instance, at what point did the stated authors of thePaxil study even become involved with the project? Were they involved by 1998, whenthe company memo acknowledging the problems with the study was written, or did theyjoin the project after the actual data was collected? If they were involved by 1998, andwere aware of the results, then how do they defend their involvement in a paper whosestated “target” was “to effectively manage the dissemination of these data in order to min-imize any potential negative impact”? When they published their paper, were they awareof negative data?
At the end of 1996, Pfizer submitted a request to the FDA for the approval of Zoloft as atreatment for Obsessive Compulsive Disorder (OCD) in children, and in October 1997the FDA approved the request (Center for Drug Evaluation and Research, 1997). Ac-cording to Healy, anyone reading the published literature on the use of Zoloft in childrenat that time would have learned of one suicidal act in a Zoloft-treated child in Pfizer’s stud-ies. Yet in the FDA’s evaluation of the data up to that point, there were six instances of sui-cidal acts in children taking Zoloft (Center for Drug Evaluation and Research, 1997;Healy, 2004). Why the discrepancy between the published and unpublished data? The ex-planations are interesting. In one paper by Alderman and his colleagues, the authors de-cided to only publish side effects that occurred in 10% or more of the patients, whichexcused them from reporting the fact that 9% of the children on Zoloft in their study hadengaged in a suicidal act (Alderman, Wolkow, Chung, & Johnston, 1998; Healy, 2003a).
In August of 2003, a major paper was published in The Journal of the American Medical
, stating that Zoloft “is an effective and well-tolerated short-term treatment forchildren and adolescents with Major Depressive Disorder” (Wagner et al., 2003, p. 1033).
For this article, the authors combined the results of two smaller studies and collectivelypresented the findings. Karen Wagner, the lead author, was also an author of some of thePaxil and Prozac studies in children.
It might seem odd that a company-sponsored paper would combine two studies into one
paper. Publishing two positive papers would certainly seem to make a stronger case for the
04_EHPP_8(1)_leo.qxd 1/27/06 3:08 PM Page 35
drug. After all, the headline, “Two Studies Show . . . ,” or better yet, “Multiple StudiesShow . . .” sounds much more convincing than “One Study Has Shown . . .” So why didthe authors combine the two studies? It likely had something to do with the fact that, asstand-alone studies, one was a failure and one showed only a trend. Only by combiningthe two studies as one could statistical significance be reached.
Just as in the studies with Paxil and Prozac, doctors trying to weigh the pros and cons of
prescribing these drugs were not given all the information in the medical literature; onlyby reading government reports (Committee on Safety of Medicines, 2004) would doctorshave known that, individually, each trial was a failure. The headline could just as easilyhave been “Two Studies Show That Zoloft Should Not Be Used in Children.”
After combining the two studies, of the children in the Zoloft group, 69% had a suc-
cessful response, whereas 59% of the children in the placebo group experienced a success-ful response: not a large difference, but enough to make the headlines. Clearly, the marketingdepartment was successful in its management of the negative information surrounding thisstudy.
There is also an interesting nuance to their experimental design. The authors do not
mention how many of the patients in this study had a history of psychiatric medication, butsince they were outpatients, probably most of them had a history of SSRI treatment. Oneof the exclusion criteria for entry into this study was a “history of failure to respond to a clin-ically adequate dosing regiment of an SSRI.” In other words, those who have not respondedto prior SSRI treatment were excluded, but it seems that another way to word this wouldbe: “Our study, designed to determine if SSRI treatment successfully treats depression, willonly use those children who have been successfully treated with an SSRI for depression.”In addition, 17 of the Zoloft-treated children, compared to 5 placebo-treated children,dropped out because of serious adverse events.
According to the FDA report, in these two studies there were six suicide events in the
Zoloft group and two in the placebo group. However, in the published paper it is not quiteas clear. Regarding suicides, in one part of the paper the authors reported, “the number ofsuicide attempts was the same in each treatment group (2 for sertraline [Zoloft] and 2 forplacebo)” (p. 1039). At least, this was the quote the press picked up on (Burton & Callahan,2003). However, in another part of the paper, the authors also mention that in terms ofsuicidal ideation there were three events in the Zoloft group versus none in the placebogroup. In addition, one Zoloft patient had an “aggressive reaction.” This description ofevents did not make the press, or receive much attention from the authors, but it did receiveattention from the regulators.
One defense of the paper’s roundabout way of discussing the suicides might be that
when the studies were being conducted, the suicide issue was not on the radar so it was over-looked. Only with the benefit of hindsight, once the issue of suicidality was raised, did aretrospective look at this study point to increased suicidality. The problem with this de-fense is that, by the time the Zoloft study was published, there were numerous hints abouta link between the SSRIs and suicides in children. Even the authors acknowledge in thepaper that regulators in the United States and United Kingdom are looking at the suicideissue (p. 1038), yet they still sidestep the suicide data in their own study.
It is surprising that the peer reviewers of the Zoloft study did not think to ask more ques-
1. Why not mention the findings of each study individually in the paper?2. What about the discrepancy between the withdrawal rates of the Zoloft and placebo groups?3. What about only a 10% difference between the drug and placebo groups?
04_EHPP_8(1)_leo.qxd 1/27/06 3:08 PM Page 36
One of the investigators involved with the Zoloft study, who was not an author and, ap-
parently, not privy to drafts of the paper, eventually wrote a letter to JAMA after the studywas published, pointing out some of the very concerns that the reviewers should have spot-ted. Though the letter unfortunately went unpublished, it stated that the conclusions ofthe paper should have been: “Sertraline is ineffective when compared to placebo and is as-sociated with increased adverse events” (Garland, 2004).
THE AMERICAN COLLEGE OF NEUROPSYCHOPHARMACOLOGY’S
As the child psychiatry profession was in the midst of a public relations problem, theAmerican College of Neuropsychopharmacology published its Preliminary Report of theTask Force on Suicidal Behavior in Youth
. The task force members claimed that, “there is suf-ficient evidence to conclude that, overall, SSRIs are effective in treating depression inchildren and adolescents” (American College of Neuropsychopharmacology, 2004, p. 4).
The task force was composed of ten members who were pivotal in the medical commu-nity’s acceptance of treating young children with these drugs. Three of the members, Gra-ham Emslie, Karen Wagner, and Neal Ryan, were authors of the Prozac, Zoloft, and Paxilstudies mentioned above. In light of the fact that the papers these authors wrote are thesource of all the confusion, it is difficult to know how to characterize their following state-ment, “all data held by FDA or pharmaceutical companies should be made rapidly avail-able to allow ACNP and other research organizations to conduct an independent evaluationof the risks and benefits of SSRIs” (p. 11). Their call for openness seems hollow when oneconsiders the way the results from their own papers were reported.
The ANCP has declared that, based on their analysis of confidential data, there is no
link between suicide and SSRIs; and while they cannot share all the data with the scientificcommunity, because of their credentials, we should trust their analysis—conflicts of inter-est aside. Yet, this is not the way science functions.
TWO PAPERS FROM THE UNITED KINGDOM
In the spring of 2004, two papers, which brought worldwide attention to the issue, werepublished in British medical journals, not American journals. This turn of events was hardlysurprising since at every step of the way the British medical community had shown moreskepticism. The first paper, published in the April 2004 issue of The British Medical Journal
(BMJ), was a review of the SSRIs for children by Jon Jureidini and his colleagues (Jureidiniet al., 2004). They concluded that the authors of the original papers had exaggerated thebenefits and downplayed the risks. The paper, whose importance cannot be underesti-mated, received a tremendous amount of media coverage. Yet, Jureidini did not really un-cover any new data, nor did he find some “smoking-gun” hidden away in a drug companybasement. He and his coauthors summarized the data that the original authors had men-tioned in their studies, and their message was clear: Before the use of these drugs becomestoo entrenched, it is time to pause and take another look at the scientific evidence. For in-stance, the original authors of the Zoloft paper thought that a 10% difference between theZoloft and placebo group was significant, but, as the Jureidini paper pointed out, or reminded
04_EHPP_8(1)_leo.qxd 1/27/06 3:08 PM Page 37
people, this statistic is not very clinically significant. The problem for academic medicineis that when the entire issue is approached as a rational scientific enterprise, where thedata drives the treatment decisions—rather than what appears to have been occurringwhere the decision to prescribe these drugs occurred well before the data was even pub-lished—then it is clear that there is little justification for the current prescribing patterns.
Two weeks later, The Lancet
published a review by Whittington and colleagues that
brought even more attention to the controversy (Whittington et al., 2004). These authorsshowed that once the unpublished literature is included in the risk-benefit analysis of thesedrugs, the benefits do not outweigh the risks. With the publication of the Jureidini andWhittington papers in mainstream medical journals, the SSRIs were in more trouble. Theeditorial in The Lancet
following the Whittington paper could not have been clearer:
That such an event [suicide] could be precipitated by a supposedly beneficial drug is a cat-astrophe. The idea of that drug’s use being based on selective reporting of favourable re-search should be unimaginable . . . (Editors, 2004, p. 1335)
What seems to bother The Lancet
editors, and others, is that, as a body of research, there
seems to be a systematic bias toward downplaying the suicide issue. Whether there is a linkbetween suicide and the antidepressants, and whether there is significant benefit to the an-tidepressants are still valid scientific questions, which will eventually be answered. However,so far, the manner in which the academic community has handled the data calls into ques-tion who will now investigate these issues. The next time one of these papers is published,the logical question on many people’s minds will be: What about the unpublished
Of course, the drug companies and authors of these studies will argue about the data men-
tioned in this paper (Clary, 2004), but, whatever the arguments, the data were incrimi-nating enough to convince two governments to intercede; the Americans with a warninglabel, and the British with an outright ban on several of the medications for people under18 years of age.
However, besides concerns about suicide, there is a very good reason for not taking the
drugs: They simply don’t work very well. Even in the adult studies, the difference betweenthe percentages of favorable responses in the drug-treated group compared to favorable re-sponses in the placebo group is minimal (Kirsch, Moore, Scoboria, & Nicholls, 2002). Inthe risk-benefit analysis, given such little benefit, there is little justification for putting upwith the suicide risk. If the studies had shown significant benefits, the suicide risk wouldhave been better tolerated by the regulatory agencies.
In addition to journal articles, textbook presentations are another area where clinicianshave apparently not been given all the information that they need to make a decisionabout suicidality and the SSRIs. During the 1994 Wesbecker v. Eli Lilly
trial, Gary Tollef-son, the head of neuroscience at Eli Lilly and a co-author of a textbook chapter on theSSRIs, stated, “there is absolutely no medically sound evidence of an association betweenany antidepressant medicine, including Prozac, and the induction of suicidal ideation orviolence” (Zuckoff, 2000). But since the Wesbecker
trial, several articles in newspapers andmedical journals have suggested that there was information available to Tollefson when hemade this statement.
04_EHPP_8(1)_leo.qxd 1/27/06 3:08 PM Page 38
In 2005, an article in the BMJ by Jeanne Lenzer discussed recent documents supposedly
showing that Eli Lilly knew about a potential link between suicidality and the SSRIs asearly as the 1980s, and that these documents were kept from the plaintiffs in the Wesbecker
trial (Lenzer, 2005). According to the BMJ article, an internal report of November 8, 1988,found that in clinical trials 38% of fluoxetine-treated patients reported new activationcompared to only 19% of the placebo-treated patients. The documents, written by employ-ees at Lilly, and pointing to a potential link between suicide and Prozac, now appear to castdoubt on Tollefson’s testimony during the Wesbecker
Lilly was not pleased with the BMJ article and disputed the claim that pertinent docu-
ments were withheld from the plaintiffs. The BMJ subsequently issued a retraction and statedthat the pertinent documents had been made available to the lawyers. However, the BMJ’sapology did not discuss the veracity of the information contained in the documents (Edi-tors, 2005).
In 2003, The New York Times
raised the issue of Tollefson’s candor at the 1991 FDA hear-
ing on the link between SSRIs and suicide (Harris, 2003). As part of the approval processfor a new medication, the FDA requires the company making such a request to release anyinformation about the status of the approval process in other countries. According to TheTimes
, under questioning from the FDA, Tollefson, who was a Lilly consultant at that point,did not mention that in 1987 the German regulators believed there was strong evidencesuggesting that there was a link between Prozac and suicide. During a subsequent trial, hewas asked about not mentioning this to the FDA and replied: “That was not the questionthat was asked so I did not answer that question.”
In 2000, The Boston Globe
cited numerous documents in Eli Lilly’s possession prior to
1998 that suggested that there was a link between suicide and Prozac. Besides mentioningthe issue with German regulators, The Globe
also pointed out that:
Internal documents show that in 1990, Lilly scientists were pressured by company execu-tives to alter records on physician experiences with Prozac, changing suicide attempts to“overdose” and suicidal thoughts to “depression.”
A lack of forthrightness might be considered acceptable in the courtroom or in govern-
ment hearings; this is not the issue, at least in this article. In the courtroom most peoplewould not offer evidence unless they are asked about it. However, even though Lilly em-ployees made the documents available to the FDA and the lawyers, it appears that the in-formation was not given to the academic community.
Putting the court proceedings aside, in 1998 the American Psychiatric Press published
its Textbook of Psychopharmacology
, which has a chapter on the SSRIs written by GaryTollefson and Jerrold Rosenbaum, who was a member of Lilly’s premarketing advisory com-mittee for Prozac. Regarding a potential link between serotonin and suicide, Tollefson andRosenbaum have a brief discussion, and then in summary cite the findings from the Amer-ican College of Neuropsychopharmacology, “that no evidence indicated that SSRIs trig-gered emergent suicidal ideation above base rates associated with depression” (Tollefson &Rosenbaum, 1998).
Like Tollefson’s statements during the Wesbecker trial, in his textbook chapter there is
no mention about any of these documents from the early 1990s, nor the findings from theGerman regulators, nor the concerns of other Lilly departments. Certainly, one possibilityis that the different departments at Lilly were not communicating, and that Tollefson andRosenbaum never discussed the suicide issue with those employees at Lilly that did haveconcerns. However, in the Forsyth vs. Eli Lilly
trial, Tollefson acknowledged that it was his
04_EHPP_8(1)_leo.qxd 1/27/06 3:08 PM Page 39
access to the Lilly database that convinced him that there was no link between suicide andthe SSRIs.
To further complicate the issue, in 1998—the very same year that Tollefson and Rosen-
baum’s chapter was published—Eli Lilly was in the process of paying $20 million for thepatent on the R(-) isomer of fluoxetine, a more specific version of Prozac. The wording ofthe patent, “furthermore, fluoxetine produces a state of inner restlessness (akathisia), whichis one of its more significant side effects” (Garnett, 2000).
In light of what was left out of the Tollefson and Rosenbaum chapter it is evident that
clinicians need to be much more skeptical when reading medical textbooks. At this pointin time, left unanswered for the academic community is: What is the purpose of a medicaltextbook? Most readers of a textbook chapter do not expect that textbook authors arewithholding information because they weren’t asked about it.
WHAT ARE CLINICIANS SUPPOSED TO THINK?
The problem for academic medicine is that the average doctor, who wants to make anevidence-based decision about the use of a certain treatment, should be able to turn to ex-periments published in the medical literature for guidance. Unfortunately, in the case ofusing psychotropic drugs in children, this is not possible.
Many people want to blame the FDA for the approval of antidepressant use in children,
or for not withdrawing the drugs from the market sooner, but out of seven requests frompharmaceutical companies for FDA approval, the FDA rejected six drugs; the only anti-depressant they approved for children was Prozac. But, while the FDA rejected six studies,the rest of the medical community accepted them. Another way of looking at this entireaffair is that the FDA had to actively intercede only because of a systemic problem in aca-demic medicine. The FDA did not just have to deal with the problematic side effect of amedication, it had to deal with a profession in denial. If academic medicine had beenworking correctly, the FDA’s drastic step would not have been necessary. Not only has thepsychiatric profession fought the FDA every step of the way—for the most part, it has stillnot acknowledged the problem (Koplewicz, 2004). Many of the SSRI defenders point outthat in the clinical trials none of the children with suicidal thoughts actually committedsuicide. However, it seems counterintuitive for these defenders to downplay the relevanceof suicidal thoughts, when it is the presence of suicidal thoughts that supposedly justifiedwriting the prescriptions in the first place.
Doctors accepted these drugs and prescribed them so willingly not because they were
given free pens, free dinners, or even free trips. They put their trust in the drugs becausethey were backed by papers that were: written by professors at the major academic medicalcenters; approved by peer reviewers; published in the major medical journals; cited in re-view articles; discussed at meetings; defended by The American College of Neuropsy-chopharmacology; and, underlying everything, given NIMH’s stamp of approval (NationalInstitute of Mental Health, 2001). In short, these papers were the talk of the town. You canhardly fault the average physician for putting faith in them.
By now, virtually everyone involved in medicine is aware of the need to take a healthy
dose of skepticism before reading pharmaceutical company brochures or when attendingtalks by pharmaceutical representatives. It is now evident that those doses of skepticism needto be taken, or maybe even increased, when reading academic publications.
04_EHPP_8(1)_leo.qxd 1/27/06 3:08 PM Page 40
ABC. (2004, December 9). Drug maker withheld paxil safety data. Prime Time Live
cast]. Retrieved September 2005 from http://a.abcnews.com/Health/story?id=311956&page=1.
Alderman, J., Wolkow, R., Chung, M., & Johnston, H. (1998). Sertraline treatment of children and
adolescents with obsessive-compulsive disorder or depression: Pharmacokinetics, tolerability,and efficacy. Journal of Child and Adolescent Psychiatry, 37
American College of Neuropsychopharmacology. (2004). Executive summary: Preliminary report of
the task force on SSRIs and suicidal behavior in youth
. Available at www.acnp.org/exec_summary.pdf
Burton, T., & Callahan, P. (2003, August 27). Antidepressant use for kids gains support. Wall Street
Center for Drug Evaluation and Research. (1997). Approval package for NDA 19839/S-017 and S-018.
Medical review of sertraline hydrochloride
. Washington, DC: Author.
Center for Drug Evaluation and Research. (2001). Application 18-936-SE-064. Medical review of flu-
. Washington, DC: Author.
Center for Drug Evaluation and Research. (2002). Application number 18-936/SE5-064. Administra-
tive documents for fluoxetine
. Washington, DC: Author.
Clary, C. M. (2004). Sertraline pediatric depression trial met highest standards of study design and
reporting. British Medical Journal, 328
Committee on Safety of Medicines. (2004). Selective serotonin reuptake inhibitors (SSRIs)—overview
of regulatory status and CSM advice relating to major depressive disorder (MDD) in children andadolescents.
Retrieved March 2004 from medicines.mhra.gov.uk/ourwork/monitorsafequalmed/safetymessages/ssrioverviewclintrialdata_101203.htm.
Dubitsky, C. (2004). Review and evaluation of clinical data placebo-controlled antidepressant studies in pe-
. Food and Drug Administration. Retrieved July 2005 from http://www.fda.gov/ohrms/dockets/ac/04/briefing/2004-4065b1-08-TAB06-Dubitsky-Review.pdf.
Editors. (2004). Depressing research. The Lancet, 363
Editors. (2005). Eli Lilly: Correction and apology. British Medical Journal, 330
Emslie, G. J., Heiligenstein, J., Wagner, K. D., Hoog, S. L., Ernest, B. A., Brown, E., et al. (2002).
Fluoxetine for acute treatment of depression in children and adolescents: A placebo-controlled,randomized clinical trial. Journal of American Academy of Child and Adolescent Psychiatry, 41
Emslie, G. J., Rush, A. J., Weinberg, W. A., Kowatch, R. A., Hughes, C. W., Carmody, T., et al.
(1997). A double-blind, randomized, placebo-controlled trial of fluoxetine in children and ado-lescents with depression. Archives of General Psychiatry, 54
Garland, J. (2004). Letter to editor
. Unpublished manuscript. Retrieved July 2005 from http://www.ahrp
Garnett, L. (2000, May 7). Prozac revisited: As drug gets remade, concerns about suicide resurfaces.
GlaxoSmithKline. (1998). Seraxat/Paxil adolescent depression: Position piece on the phase III clinical stud-
Grinfeld, J. (1998, March). Psychoactive medications and kids: New initiatives launched. Psychiatric
Harris, G. (2003, August 7). Debate resumes on the safety of depressions wonder drugs. The New
. Retrieved September 2005 from http://query.nytimes.com/gst/fullpage.html?res=9E0DE4DD1731F934A3575BC0A9659C8B63&sec=health&pagewanted=print.
Healy, D. (2003a). Let them eat Prozac
. Toronto, Canada: James Lorimer & Company.
Healy, D. (2003b). Lines of evidence on the risks of suicide with selective serotonin reuptake in-
hibitors. Psychotherapy and Psychosomatics, 72
Healy, D. (2004). Suicidal evidence not addressed by FDA
. [Letter from David Healy to Peter J. Pitts,
Associate Commissioner for external relations, Food and Drug Administration]. Unpublishedmanuscript. Retrieved July 2005 from http://www.ahrp.org/risks/healy/FDA0204.php.
04_EHPP_8(1)_leo.qxd 1/27/06 3:08 PM Page 41
Jureidini, J. N., Doecke, C. J., Mansfield, P. R., Haby, M. M., Menkes, D. B., & Tonkin, A. L.
(2004). Efficacy and safety of antidepressants for children and adolescents. British Medical Jour-nal, 328
Jureidini, J. N., & Tonkin, A. L. (2003). Paroxetine in major depression. Journal of American Acad-
emy of Child and Adolescent Psychiatry, 42
Keller, M. B., Ryan, N. D., Strober, M., Klein, R., Kutcher, S. P., Birmaher, B., et al. (2001). Effi-
cacy of paroxetine in the treatment of adolescent major depression: A randomized, controlledtrial. Journal of American Academy of Child and Adolescent Psychiatry, 40
Kirsch, I., Moore, T. J., Scoboria, A., & Nicholls, S. S. (2002). The Emperor’s new drugs: An analy-
sis of antidepressant medication data submitted to the US Food and Drug Administration.Prevention and Treatment
. Retrieved July 2004 from http://www.alpha-stim.com/Information/Technology/Research/Research_PDF/EmperorsNewDrugs.pdf.
Koplewicz, H. (1997). It’s nobody’s fault: New hope and help for difficult children and their parents
Koplewicz, H. (2004). Misguided resistance to appropriate treatment for adolescent depression. Jour-
nal of Child and Adolescent Psychopharmacology, 14
Laughren, T. P. (2004, January 5). Background comments for February 2, 2004 meeting of the Psy-
chopharmacological drugs advisory committee (PDAC) and pediatric subcommittee of the anti-infectivedrugs advisory committee
Lenzer, J. (2005). FDA to review “missing” drug company documents. British Medical Journal, 330
Leo, J. (2004). NIMH guilty too? Psychiatric News, 39
Mosholder, A. D. (2004). Executive summary: Suicidality in pediatric clinical trials with paroxetine and
other antidepressant drugs
. Follow up to 9-4-03 consult. Retrieved July 2005 from http://psy-chrights.org/Research/Digest/AntiDepressants/Mosholder/MosholderReport.pdf.
National Institute of Mental Health. (2001). Let’s talk about depression
. NIH publication No. 01-
Tollefson, G. D., & Rosenbaum, J. F. (1998). Selective serotonin reuptake inhibitors. In A. F.
Schatzberg & C. B. Nemeroff (Eds.), The American psychiatric press textbook of psychopharmacol-ogy
(2nd ed., pp. 219–238). Washington, DC: The American Psychiatric Press.
Wagner, K. D., Ambrosini, P., Rynn, M., Wohleberg, C., Greenbaum, M. S., Childress, A., et al.
(2003). Efficacy of sertraline in the treatment of children and adolescents with major depressivedisorder: Two randomized controlled trials. Journal of the American Medical Association, 290
Whittington, C. J., Kendall, T., Fonagy, P., Cottrell, D., Cotgrove, A., & Boddington, E. (2004). Se-
lective serotonin reuptake inhibitors in childhood depression: Systematic review of publishedversus unpublished data. The Lancet, 363
Zuckoff, M. (2000, June 8). Prozac data was kept from trial, suit says. The Boston Globe
, p. 1.
Requests for offprints should be directed to Jonathan Leo, PhD, 17607 White Tail Court,
Parrish, FL 34219. E-mail: email@example.com
04_EHPP_8(1)_leo.qxd 1/27/06 3:08 PM Page 42
PLASTIC LITHIUM CELL HOLDERS “1/2AA” • “CR2” • “2/3A” FEATURES APPLICATIONS • Snap-In PC contact holds in position for wave soldering • Computer memory, power transfer and back-up systems • Tin Nickel Plated contacts for excellent solderability and durability • Video and telecommunications power back-up requirements • Polarity cle
MEDICAL INFORMATION FOR NATURE CAMP—2013 Please fill out all four pages and return within four weeks of the start of camper’s session to If possible, please arrange for camper to have examination by physician no more than four weeks prior to the start of his or her session so that medical information will be as up to date as possible. The American Camp Association recommends tha