WASHINGTON, DC— Physicians who treat children with depression are left with many unanswered questions in the wake of the FDA’s recent request that a black box warning be put on the labels of all antidepressant medications. As part of its multipronged public health advisory issued in October, the agency directed drug manufacturers to add the more prominent warning regarding an increased risk in suicidal thoughts and behavior for children and adolescents taking antidepressants, and it also emphasized the need for close monitoring of patients who begin these medications. The FDA is also compiling a Patient Medication Guide that will be given to patients who use the drugs to advise them of this risk as well as precautions that can be taken. The new language does not prohibit the use of antidepressants in children and adolescents but rather, in essence, may shift more focus and responsibility to health care professionals as far as assessing the risks and benefits of the medications they prescribe for their patients.

“The present issue for pharmacotherapy is that the risk/benefit ratio for pediatric patients remains favorable with fluoxetine, [which] will likely be the only antidepressant in the near future with an indication for pediatric major depressive disorder,” said Bruce Waslick, MD, who is an Associate Clinical Professor of Psychiatry and Director of the Children’s Day Unit at the Columbia University Medical Center in New York City. “There’s really no clear guideline to something that we are thinking a lot about—What do we do if somebody does not do well on fluoxetine? What are we going to do as the next step? Do we [prescribe] another selective serotonin reuptake inhibitor (SSRI), even though there is no robust efficacy, no indication, plus the black box warning? Do we add things to SSRIs?”

Complicating matters is the inherent difficulty in trying to differentiate between children who develop suicidal thoughts as a result of medication and those who become suicidal due to their depression. “If you are treating depressed kids, you are going to see some getting suicidal,” Dr. Waslick continued. “It’s not clear to me how you are going to identify the kids who are having an adverse medication reaction versus kids who are getting suicidal during the course of their depression. And there’s nothing given from the data that would lead you to suggest that you can actually distinguish those two things. If kids get suicidal and they are on an antidepressant, should we wait longer and see if they will get better from the antidepressant, or should we get them off because we have to assume that they are all adverse events? It’s not clear to me how we’re going to be able to do that. We’re going to get some experience with it over the next few years.” Dr. Waslick made his comments during a presentation at the 51st Annual Meeting of the American Academy of Child and Adolescent Psychiatry.

Potential medical-legal liability that may arise from a child who has an adverse event is a cloudy issue at the moment. Considering the fact that, according to the FDA, on average there is one adverse event for every 50 patients treated with SSRIs, many clinicians will eventually encounter such a case. “Does the black box warning protect the company but hurt the physician?” asked Dr. Waslick. “Does it help the physician if you institute the monitoring and some bad thing happens? Are [physicians] going to be more or less liable? I’m not sure what medical-legal issues are going to result from this. It makes me nervous. It makes me cautious and conservative. I’m not sure what the physician liability responsibilities are here.”

REVIEW OF EFFICACY DATA

The leading medical journals are lagging behind in terms of published efficacy data concerning antidepressants, with the FDA basing its actions in the past year largely on unpublished data, according to Dr. Waslick. In looking at the published efficacy data that are available, tricyclic antidepressants, for example, have been shown to have a “reasonably robust positive effect” in adults, Dr. Waslick noted. They also have been studied in children and adolescents with depression in sample sizes ranging from nine to 60. In comparison, some of the SSRI studies that were in the FDA pool were as large as 220 subjects. With tricyclic antidepressants, “one of the real compromises with all the studies that have been reviewed in terms of efficacy and safety is that we really only have short-term outcomes,” said Dr. Waslick. “No single study ever demonstrated any efficacy of a tricyclic above and beyond what you would achieve with placebo. Even if you put together these studies in a meta-analysis, the meta-analysis fails to support the superiority of tricyclic antidepressants compared with placebo in pediatric patients. So the tricyclic antidepressants didn’t seem to be a particularly promising treatment for kids.”

Thus far, there have been six published studies concerning efficacy of SSRIs, as well as one published serotonin norepinephrine reuptake inhibitor (SNRI) study, which was negative. Simeon et al found that children who took fluoxetine did not appear to improve more than did those who took placebo. The other five studies were all reported as positive, beginning with Emslie and colleagues’ 1997 trial with fluoxetine, which found that kids who took it did better than did those who got placebo in a short-term time frame.

Keller and colleagues’ report in 2001 may have sparked the first concerns about suicidality in children being treated with SSRIs. This large, three-arm study randomly assigned adolescents to short-term treatment with paroxetine, imipramine, or placebo. The trial was designed “to take the two active medications and compare them with placebo, not really to compare the two active arms,” Dr. Waslick pointed out. “There were significant positive effects for paroxetine compared with placebo. Imipramine really showed no difference from placebo in any of the primary outcome measures. [Keller] reported that paroxetine was well tolerated, although there were high placebo rates. If you read this article closely, you actually find that the two primary outcome measures that the investigators had a priori decided would be the test of efficacy were actually not significantly different from placebo in the paroxetine group. There are a lot of secondary measures, supplementary measures. But the two primary outcome measures looked like they were moving in the right direction, but they weren’t significantly different than placebo. That becomes important when you hear what the FDA looks at in terms of this study.”

What caught people’s attention in Keller’s study was what the investigators referred to under an adverse events heading as emotional lability, along with a parenthetical reference to suicidality, suicide ideation, and suicidal experiences. “There were higher rates of what they called emotional lability in the kids who got paroxetine compared to the placebo group,” commented Dr. Waslick. “Some of that emotional lability was a larger heading for issues regarding suicidality. Some people seemed to notice that and got the ball rolling on this.”

In another multisite study with fluoxetine involving 219 children and adolescents, Emslie et al found overall response rates of 65% in the fluoxetine group and 53% in those taking placebo. Remission rates were 41% in the fluoxetine group, and 20% for placebo, a number “somewhat compromised by the 38% drop-out rate in the placebo arm,” noted Dr. Waslick. “There were no indications in this particular trial of risks of suicidality. Overall, [it was] considered a positive study.... So, if we summarized the five published positive SSRI trials, one was supported by nonprivate funding, four were supported by private industry, and the total [number of subjects] of the published trials is about 1,100 by my calculation. There was mild to moderate support for the short-term efficacy of SSRIs, namely fluoxetine, paroxetine, sertraline, and citalopram.”

THE TURNING TIDE

About a year ago, when the FDA announced that it was reviewing data concerning antidepressant drugs from about 24 studies and 4,400 patients, “I realized that there was a whole slew of unpublished data that the FDA was actually looking at, which was a warning,” said Dr. Waslick. At about the same time, the Medicines and Healthcare Products Regulatory Agency (MHPRA)—the UK equivalent of the FDA—issued summary findings of its own after systematically analyzing each antidepressant medication’s development program. “The MHPRA found that there was no evidence for efficacy for any SSRIs, and no efficacy for any of the SNRIs except for fluoxetine,” stated Dr. Waslick. “It also concluded that the adverse events rates exceeded placebo in most of the medications, which is not really that surprising.”

The MHPRA had specific concerns about suicide-related events in a variety of the medications and determined that the risk/benefit ratio was unfavorable for the entire class of drugs except for fluoxetine in children and adolescents, which led it to issue a contraindication in the UK. At the same time, the UK agency also found no deaths in any of the trials, and “that’s been consistent throughout the MHPRA and the FDA work,” said Dr. Waslick. “So there’s a confusion between the published and unpublished data. Only six of the 15 pediatric depression trials available to the MHPRA have made it to peer-reviewed publication as of October 2004. Five of the six published trials are positive. Negative trials tend to go unpublished.... The MHPRA, and I think the FDA has followed suit, concluded that two of the positive published trials are really not positive by their criteria of a priori–defined primary outcome measures. So they don’t consider the Keller study [and others] robust evidence of efficacy enough to say that these are really positive studies. There are hints of positive signals, but they are not robustly positive. That gets confusing, because what is coming out in the published literature is not being accepted by regulatory agencies as positive data.... When we are out there trying to make evidence-based treatment recommendations, we tend to rely too much on the published literature, but we have to remember that the published literature may be biased in terms of positive results.”

In summing up the overall efficacy data, “there really is, at least in the data from the drug company industry, underwhelming evidence of efficacy of most of the antidepressants in children and adolescents outside of fluoxetine,” affirmed Dr. Waslick. “At least the industry-sponsored studies have exhibited, if not been compromised by, relatively high placebo rates in the clinical trial data. It’s very hard to show that your medication or treatment works if you are getting placebo response rates of 60% to 65%. And you might make very different conclusions if you relied only on the published literature [rather] than combining the published literature with the unpublished literature.”

FDA TAKES AIM

In February 2004, the FDA held its first public hearing regarding antidepressant use in pediatric populations. The hearing was used to gather information and to discuss with the agency’s advisory boards ways in which available data could be analyzed. In March, the FDA issued a general warning about antidepressant drugs being associated with, but not causing, a risk in suicide. “It strengthened concerns about suicidality developing during treatment that had already been in the labeling for all the antidepressants,” stated Dr. Waslick. “It didn’t really make it specific to kids or to any type of illness or indication—a general warning for monitoring your patients closely for worsening of depression or development of suicidality. It didn’t talk about medications causing suicidality; it just really encouraged clinicians to warn. In the meantime, they were setting up a second analytic strategy, which was to take a look at some of the adverse events data, reclassified as to the presence or absence of suicidality using a more expert classification system.”

In September, after the reanalysis was completed, the FDA held a second public hearing. Members of an FDA advisory committee voted 15 to eight in favor of recommending a black box warning, and on October 15 the FDA ultimately issued its strongest possible warning on the issue. “The conclusions that came out of the September FDA hearing, which has really set off the chain of events leading to where we are today, is that the FDA analysis, using the classifications from the Columbia University–led investigators, ... concluded that there is a low- magnitude, low-frequency, suicide signal in the adverse events and serious adverse events data set,” Dr. Waslick commented. “And [the FDA] feels at this point that it can sort of make definitive conclusions that there is a causal link—SSRIs, second-generation antidepressants are causally linked to suicidality. I thought it would be more of an association, but [the FDA] is talking about causal links. At the same time, evaluation of systematically collected data on depression rating scales [showed] ... there really was no signal. The adverse events, which are more spontaneously collected, versus systematically collected data, are a little bit at odds here, but they really represent different things.”

The FDA also concluded that there appeared to be a class effect among all drugs regarding suicidal signals, meaning that although certain medicines seemed to have a stronger or weaker association, no analysis basically absolved any medication of a risk. “[The FDA] felt that there was a class effect, and that’s why it didn’t contraindicate or put labels on particular medications,” explained Dr. Waslick. “It basically concluded that no medicine in the class is without risk. [The FDA] is going to differentially label things, but it’s not around the risk, it’s going to be around the efficacy.”

The FDA did take into account data from the Treatment for Adolescents With Depression Study (TADS), the only study that has found an increased risk of suicidality among patients in the fluoxetine trials. Investigators who conducted TADS found that for moderately to severely depressed adolescents, a combination treatment of medicine and cognitive behavioral therapy appeared to work best. “Cognitive behavioral therapy alone did not really distinguish itself from placebo,” noted Dr. Waslick. “So the most effective treatments we have for depression include at least a component of medication. We have to remember that even though we are going to have those black box warnings.”

Dr. Waslick called attention to the fact that treatment-emergent suicidality per the FDA database included the full range, from passive suicidal thinking to serious suicide attempts. “It’s just not the kids who are having the most severe reactions,” he said. “There are kids who are having a very wide range of suicidal experiences. You can’t just say, well, it is a suicide attempt or it’s [adolescents] who start obsessively planning suicide. It could even include people who have impulsive suicidal crises.”

RECOMMENDED MONITORING GUIDELINES

In addition to issuing the warning, the FDA is also recommending that pediatric patients being treated with antidepressants for any indication should be closely monitored for clinical worsening, as well as agitation, irritability, suicidality, and unusual changes in behavior, especially during the initial few months of drug therapy or during dosage changes. The agency has advised that such observation would ideally include at least weekly face-to-face contact with patients or their family members or caregivers during the first four weeks of treatment, then visits every other week for the next four weeks, then at 12 weeks, and as clinically indicated beyond 12 weeks. “That’s a very high frequency of monitoring SSRIs,” commented Dr. Waslick. “It’s probably more closely than I monitor kids generally on SSRIs.”

Dr. Waslick also addressed concerns about any potential liability resulting from undertreating patients. He again advised that based on the TADS data, combination treatment is best for moderately to severely depressed children and that close monitoring should be done for all patients. “That’s what I am going to recommend at this point in time,” he commented. “I think it would be a great loss to mental health care if people just stop prescribing these medicines or don’t get educated and trained. I think we have to come up with more reasonable guidelines. I think the monitoring level is hard for a child psychiatrist to follow, never mind a pediatrician or a family practitioner. [The FDA] does use the word ‘ideally.’ It doesn’t mean you have to. But, ideally, you would do it.... I still think the full story is not told with these medications.”

—Colby Stong

Suggested Reading
Emslie GJ, Heiligenstein JH, Wagner KD, et al. Fluoxetine for acute treatment of depression in children and adolescents: a placebo-controlled, randomized clinical trial. J Am Acad Child Adolesc Psychiatry. 2002;41:1205-1215.
Emslie GJ, Rush AJ, Weinberg WA, et al. A double-blind, randomized, placebo-controlled trial of fluoxetine in children and adolescents with depression. Arch Gen Psychiatry. 1997;54:1031-1037.
Keller MB, Ryan ND, Strober M, et al. Efficacy of paroxetine in the treatment of adolescent major depression: a randomized, controlled trial. J Am Acad Child Adolesc Psychiatry. 2001;40:762-772.
March J, Silva S, Petrycki S, et al. Fluoxetine, cognitive-behavioral therapy, and their combination for adolescents with depression: Treatment for Adolescents With Depression Study (TADS) randomized controlled trial. JAMA. 2004;292:807-820.
Simeon JG, Dinicola VF, Ferguson HB, Copping W. Adolescent depression: a placebo-controlled fluoxetine treatment study and follow-up. Prog Neuropsychopharmacol Biol Psychiatry. 1990;14:791-795.

Return to table of contents