With increased concern about suicidal impulses in children taking antidepressants, the FDA has decided to step in. Two committees held a meeting to discuss various plans for classifying suicidal events, along with some of the difficulties in deciding whether certain events qualify as suicide attempts.
Two advisory committees of the U.S. Food and Drug Administration met in an unusual joint session in February to consider whether antidepressant treatment provokes suicidality in children. At the end of this first of two planned meetings, they recommended that the FDA warn practitioners about this possible risk.
The Psychopharmacological Drugs Advisory Committee and the Pediatric Subcommittee of the Anti-Infective Drugs Advisory Committee received summaries of spontaneous drug reports and of experiences in clinical trials with children; called expert witnesses on suicide research; and heard dramatic testimony from families of suicide victims, as well as from those whose children had benefited from antidepressant treatment.
According to Russell Katz, M.D., the events identified in the clinical studies by the pharmaceutical manufacturers will be independently reclassified by a group at Columbia University with particular expertise in adolescent suicidality. At a second meeting this summer, the committees may also hear expert testimony on improving prospective assessment and monitoring for suicidality and possibly improving antidepressant clinical trial designs to more clearly ascertain whether benefit outweighs risk.
Katz indicated that the FDA investigation of the possible emergence of suicidality with selective serotonin reuptake inhibitors or other newer antidepressants has been complicated by inconsistencies in the initial classifying and reporting of potential events and by disparate experiences in comparable studies with the same drug.
He defended the investigative process as systematic and deliberative, however, and indicated that answers and appropriate actions would be forthcoming. "We are, of course, aware that there is great concern among the families of children and adolescents with depression about whether or not these drugs can be used safely," Katz told attendees of the joint session.
"It is absolutely critical, in our view, that we make every effort to provide the best answer possible to this question," Katz said. "The wrong answer in either direction, prematurely arrived at, could have profound negative consequences for the public health."
The FDA issued alerts last October advising particularly close supervision of at-risk patients when initiating antidepressant treatment (Psychiatric Times December 2003, p27). The alerts were based on preliminary assessment of events with nine antidepressants, which had been undertaken after the FDA identified an increased risk of possibly suicide-related events in one of the three pediatric supplemental clinical studies with paroxetine (Paxil), for which it issued an advisory alert in June.
In contrast to the advisories issued by the FDA, the United Kingdom's Medicines and Healthcare Products Regulatory Agency (MHRA) acted on the paroxetine findings by issuing a directive in June and then a label change contraindicating use for depression in children. The MHRA also contraindicated venlafaxine (Effexor) after the manufacturer found a signal of suicidality in the data review requested by the FDA and issued its own "Dear Doctor" warning letter in August.
The MHRA then acted on the preliminary FDA review of nine antidepressants, announcing in December that sertraline (Zoloft), citalopram (Celexa) and escitalopram (Lexapro) are contraindicated for treatment of pediatric major depression in the United Kingdom. Other drugs with possible signals of suicidality in the FDA preliminary review (nefazadone [Serzone] and bupropion [Wellbutrin]) are not approved in the United Kingdom. Data from the one study with mirtazapine (Remeron) and the three with fluoxetine (Prozac) did not suggest the association.
Only fluoxetine remains approved for the treatment of pediatric depression both in the United States and the United Kingdom for demonstrating superiority to placebo in the pediatric supplemental studies.
(A review article published in the April issue of the British Medical Journal compared six trial studies of safety and efficacy of antidepressants for children and adolescents. The study found, "Antidepressant drugs cannot confidently be recommended as a treatment option for childhood depression"--Ed.)
The American College of Neuropsychopharmacology (ACNP) took a different stance after reviewing published and unpublished data from controlled trials of the antidepressants in children, in addition to epidemiological studies and autopsy reports. The ACNP announced in the week preceding the FDA advisory committee meeting that it had not found that SSRIs increased the risk of suicidal behavior in youth with depression.
Two members of the ACNP task force also participated in this joint meeting of FDA advisory committees: Andrew Leon, Ph.D., as a member of the psychopharmacological committee, and David Shaffer, FRCP (Lond), FRC Psych, as an invited expert on suicide in adolescents. Both noted the preliminary nature of the ACNP conclusions in their conflict-of-interest disclosure statement, however, and welcomed additional pertinent data from manufacturers and the FDA for the ACNP to evaluate.
Distinguishing Drug Reaction From Depressive Symptom
The committees sought to confirm that the potential events in children represent suicidality and then ascertain whether they emerge from or are exacerbated by antidepressant treatment, or are symptomatic of depression unrelieved by the treatment. A similar assessment was conducted by the Psychopharmacological Drugs Advisory Committee in 1991 in response to concerns that fluoxetine prompted suicidality, and analyses were performed on data from studies with fluoxetine and other antidepressants in adults. Although it was concluded that there was no increased risk of completed suicide associated with either an SSRI antidepressant or placebo in adults with major depression, several possible drug-related mechanisms were considered. These included a paradoxical worsening of depression, onset of akathisia, induction of anxiety or panic, and switching of bipolar depression to mixed state.
The current assessment of causality in children occurs against a high background of suicidality, according to Shaffer. "Suicidal ideation in high school students is extraordinarily common," he reported. "Almost 20% of American high school students will think about suicide during the past year."
Shaffer elaborated, "Not only do many adolescents attempt and think about suicide, but they do it quite often, so that from the studies that we have, about half of suicide attempters will make only one attempt a year, and nearly a half will make two or more."
In addition to the prevalence of suicidality among adolescents, there is also extensive use of antidepressants in this population, Giana Rigoni, Pharm.D., M.S., told the committees. "Approximately 10.8 million total prescriptions were dispensed for all SSRIs and atypical [antidepressants] in this population [17 years and younger], representing a substantial 7% of the market in 2002."
Solomon Iyasu, M.D., M.P.H., described the adverse drug reactions with paroxetine spontaneously reported in the year following the pediatric supplemental studies. These included nine completed suicides, 17 suicide attempts and suicidal ideation in 11 patients. There had been no completed suicides within the pediatric controlled studies with paroxetine or the other antidepressants.
Iyasu indicated that the majority of attempts were made by patients treated for major depression or bipolar disorder and that approximately one-third of these patients used concomitant medications. He noted a prior history of suicide attempts in some patients and that, in others, "there was no negative history of this."
Spontaneous reports are characterized by variability of quality and detail, Iyasu explained. Timing of the event in relation to the medication is often unclear, and there is often a lack of follow-up information and ascertainment by medical professionals.
"The causality assessment was very difficult in many of the reviews that we have done with these reports," Iyasu acknowledged. "Many of the psychiatric events that were described in the reports occurred in patients with underlying psychiatric disorders. Therefore, severity of illness/underlying disease may play a role, and it was very difficult to disentangle its effect from what might have been going on."
The potential events from the controlled studies proved no less problematic for the FDA to assess. Thomas Laughren, M.D., described inconsistencies in documenting and omissions in reporting, which had hindered subsequent analyses. Among these were: omission of cases coded as "accidental injury," although "accidental overdose" was accounted for; omission of events not deemed treatment-emergent, with inconsistent criteria across studies for classifying events as reportable; and failure to maintain blinded control in identifying reportable cases.
"One particular example," Laughren explained, "was a child who stabbed himself in the neck with a pencil while taking a test. Now, this probably was an accident, but it occurred to us that we want to see all of ... the events that had been excluded as accidental injury."
The FDA elected to contract with the Columbia University Suicidality Research Group to reclassify the events in the studies. The FDA will reassess these data for causality, subject to the limitations of the original monitoring and documentation. Kelly Posner, Ph.D., explained that the Columbia University group would apply standardized, research-supported definitions in identifying suicidal behaviors.
"We have been able to demonstrate excellent reliability utilizing these definitions and this classification system in NIMH [National Institute of Mental Health]-funded treatment, biological, and genetic trials across the life span," Posner said.
Weighing Antidepressant Benefit Against Risk
With causality yet unclear, David Antonuccio, Ph.D., and Irving Kirsch, Ph.D., posed the question of whether uncertain antidepressant drug benefits for children outweigh undetermined risk. Both researchers have published reviews and analyses of placebo-controlled trials that fail to demonstrate antidepressant efficacy (PT September 2002, pp6-9).
"Clinically meaningful benefits have not been adequately demonstrated in depressed children, therefore, no extra risk is warranted," Antonuccio told the panelists. "The highest possible standard should be applied to scientific data involving drug treatment of children, because children are essentially involuntary patients."
Laughren acknowledged in the meeting the troubling failure of all drugs except fluoxetine to demonstrate superiority to placebo in the pediatric supplemental studies. In briefing documents provided earlier to the committee members, however, Laughren argued that these results point to the need for better study design, rather than proof of lack of efficacy.
"We at the FDA ... do not view negative studies as proof of no benefit," Laughren stated. "In our view, absence of evidence for effectiveness in most of these programs does not constitute evidence of absence of benefit for these drugs in pediatric patients."
Laughren referred to methodological problems that can interfere with separating antidepressant drug effect from placebo. He suggested that the committees might consider alternative designs to ascertain drug benefit and recommend the most promising ones. Laughren suggested randomized withdrawal as one possibility, in which the initial medication responders in a controlled, randomized trial are randomized again to drug or placebo and followed, in order to establish relative relapse rates.
The emphasis of the meeting on re-analysis of controlled studies was clearly unsatisfactory to many of those speaking in the public hearing session about losing loved ones to suicide. Gary Cheslek, D.D.S., a dentist from Vicksburg, Miss., told the panelists that, while they might consider the individual stories in the public hearing to be anecdotal, he considered an "anecdote" to be "the euphemism the manufacturers of Prozac, Paxil, Effexor and Zoloft use to describe the thousands of reported out-of-character, violent, homicidal and suicidal events that occur in a vulnerable subset of patients who ingest their SSRI antidepressants."
In the end, the advisory committees agreed with the FDA plan to have the expert group at Columbia University re-analyze the data from the efficacy trials, although with some concern expressed that there would be insufficient power in the data to reveal drug effects.
In addition, they recommended that the FDA issue stronger warning indications to clinicians regarding possible risks of these medications, without recommending that their use be contraindicated. (In response to these hearings, the FDA issued a Public Health Advisory on March 22. In the advisory, the FDA asked pharmaceutical manufacturers to include a warning statement recommending close watch of pediatric patients for signs of depression and suicidality--Ed.)
Matthew Rudorfer, M.D., chairperson of the Psychopharmacological Drugs Advisory Committee commented, "We think such warnings are required to elevate the level of concern and attention that practitioners use in prescribing them."