Reexamining the Link Between Antidepressantsand Suicidality in Children and Adolescents
In October 2004, the FDA mandatedthat a black box warning be addedfor all antidepressants used inchildren and adolescentsa move thatsparked widespread concern about theuse of these drugs.1 The FDA's decisionfollowed similar action by theBritish Medicine and HealthcareProducts Regulatory Agency.
The FDA based its decision on areview of pooled analyses of 9 antidepressants(selective serotonin reuptakeinhibitors [SSRIs] and other novelantidepressants) used in 24 short-term(4 weeks to 16 weeks), randomized, placebo-placebo-controlled trials, both publishedand unpublished.2 The trials involved4587 children and adolescents with majordepression, obsessive-compulsive disorder,and other disorders. The initial metaanalysisrevealed an average risk ofbroadly defined suicidal behavior of 4%as determined by adverse event reporting,which is twice the placebo rate.
The FDA also concluded thatalthough there had been trials supportingthe efficacy of citalopram, sertraline,and paroxetine, only fluoxetine hadmet the efficacy standard of showinga benefit in at least 2 randomizedcontrolled trials. They noted that a third,multi-site, randomized controlled trial,the Treatment for Adolescents withDepression Study (TADS), confirmedthat fluoxetine, alone or in conjunctionwith cognitive-behavioral therapy, was an effective treatment for depression.3In the TADS, 61% of the participantsresponded to fluoxetine alone, and 71%responded to fluoxetine plus therapy,which was twice the response rate of35% for participants who receivedplacebo. Weighing the efficacy dataagainst the risks of suicidal behavior andideation, the FDA concluded that onlyfluoxetine showed significant advantageover placebo and, therefore, thebenefit of this drug exceeded the risk.
Hammad and colleagues4 reported ona more detailed reanalysis of all therandomized placebo-controlled studies,including the TADS data. In this study,the investigators analyzed the suicidal behavior data separately from otheradverse events; they also separated thedata on major depressive disorder(MDD) trials from data from other disorders.The overall risk ratio of suicidalideation or nonfatal self-injurious eventscompared with placebo was found to be1.95 (95% confidence interval [CI], 1.28to 2.98%) for all disorders and 1.66 (95%CI, 1.02 to 2.68) for MDD.
Table 1 shows the risk ratios forsuicide attempts and suicide behaviorfor each drug within MDD trials ascompared with all other trials. Althoughthere were some differences in thedegree of risk among the 9 antidepressantsevaluated, the FDA did notexempt any antidepressant from theblack box warning because of differentialrisk rates. The FDA has notcontraindicated any of the medications for use in pediatric populations but encouragesphysicians to balance the riskwith clinical indications.
Reactions to the FDA warning
The FDA's actions and related studieshave caused public consternation andsparked an ongoing debate in the medicalcommunity. The profusion of commerciallysponsored trials that werenot required to be reported for analysiscreated the perception of a reportingbias.5 The FDA quickly moved tostart requiring the reporting of all datapositive and negative.
Psychotropic medication use in childrenand adolescents has come under even more scrutiny with concerns aboutdirect-to-consumer marketing and theincrease in medication prescriptions forchildren. These concerns were illustratedby the recent recommendationof an FDA advisory committee to adda black box warning and further investigatepossible cardiovascular risksof medications used to treat attentiondeficit/hyperactivity disorder. However,another FDA panel then reversed therecommendation to add a black boxwarning.
Since the FDA's first public healthadvisory on antidepressants, reportssuggest that there has been a 20%drop in the number of antidepressantprescriptions dispensed for minors.6The public health consequences ofthis are unclear. In June 2005, a reportissued by the AMA Council on Scientific Affairs stated that SSRIsshould continue to be available as acomponent of depression treatment forchildren and adolescents.7 The AMArequested an independent review of thecurrent data and asked the FDA to evaluatethe impact of its regulatory actionson treatment patterns, compliance, andaccess to care.
Significance of the findings
The significance of these data is farfrom clear, because there are a varietyof alternative explanations as well asmethodologic limitations that make thefindings concerning suicide risk andclinical efficacy difficult to interpret.For instance, the trials were carriedout in a pediatric population, in whichactual suicide is rare, so the events reportedwere very small in number. Therelationship of suicidal ideation andsuicidal behavior to completed suicidein this age group is also unclear. Suicidalideation is common among adolescents,occurring in approximately 19%.
The studies analyzed by the FDAinvolved spontaneously reported adverseevents rather than explicitlyassessed behaviors. They also lackeduniformity in how suicidal behaviorswere defined and lacked adequatelyexplicit measurements of suicidal behaviorsor ideation, including measures ofseverity of intent. It is important toemphasize that there were no suicidesamong the nearly 5000 cases reviewedby the FDA and that behaviors consideredby external consultants as probablesuicide attempts were rare. Thus, thepooled data include behaviors that maynot pose a significant risk of suicide.Finally, adherence to assigned treatmentwas not measured, and investigators didnot account for the possibility thatdiscontinuation symptoms contributedto adverse events.8
With respect to efficacy, the durationof the trials (less than 16 weeks) was,in general, too short to allow identificationof the long-term beneficial effectsof antidepressant medications. Simonand colleagues9 recently published areport on attempted and completedsuicide in a large HMO sample forwhom antidepressants were prescribed.They found that the highest risk ofsuicide attempts and completion was inthe month before antidepressant use, andthe risk went down steadily during thecourse of treatment over 6 months, evenfor patients younger than 18.
The large number of multi-site trialsincluded in the FDA meta-analysis presented additional problems for efficacymeasurement. These include high variancebecause of multiple sites with fewpatients in each site, probable variationsin methodologic and recruiting strategies,and variable placebo response rates.8
Mann and colleagues10 noted otherproblems with these trial results. Theyfound that the trials did not documentpast suicide attempts nor did they include,in general, patients with current suicidality.As a result, these authors recommendedthat future trials be designed withhigh-risk patients, using stratification toavoid confounding effects.
Various alternative explanations forthe trial results have been advanced.Among these is the possibility that thepatients receiving medications weremore able to report suicidal behavior(either because they elicited attentionfor other reasons or they were morecommunicative).4 It has also beenposited that some of the adverse effectsof antidepressant treatmentsuch asinsomnia, agitation, restlessness, andirritabilityoccurred in patients at riskfor bipolar disorder who may not havebeen adequately monitored or treated.11
Epidemiologic data also fail to supportthe conclusions of the FDA meta-analysis,since the recent increase in SSRIprescriptions has been correlated with adecrease in the youth suicide rate.12 Adultdata also show a decrease in the suiciderate corresponding to the period of increasedSSRI prescriptions (with theexception of the elderly population).
In addition, autopsy studies havefailed to find antidepressant use in mostsuicide victims, which is also inconsistentwith the FDA's findings.4 A Swedishtoxicology study compared suicides withnatural deaths and reported that no SSRIswere detected in 52 suicides in personsyounger than 15 years. In the 15 to 19year-old age group, persons withdetectable concentrations of SSRIs hada lower relative risk of suicide overa period of 9 years than those withdetectable levels of other antidepressants.13 Data on regional trends inprescribing antidepressants correlatedwith suicide statistics suggest that antidepressanttreatment is inversely relatedto suicide for older adolescents and formales in general, who are at highest riskfor suicide.14
Despite these caveats, the available dataappear to warrant the conclusion thatthere is a modest increase in suicidalideation and suicidal behavior associatedwith antidepressant use in childrenand adolescents. Since there were nooccurrences of suicide in any of thetrials, a far larger sample size would beneeded to demonstrate increased riskof actual suicide.
The more complicated issue is howto apply this information in clinical practice.While it does make sense to becautious in using antidepressants asfirst-line treatment for mild to moderatemajor depression, especially in theabsence of a trial of psychotherapy,caution needs to be weighed against therisk of limiting treatment options fordepressive disorders that can be associatedwith major psychosocial andbiologic impairments.
Observing that much of the increasedprescribing of SSRIs has occurred inpatients with mild depression andanxiety, Gunnell and Ashby15 posit thatthe risk-benefit ratio may depend on theindividual's underlying suicide risk. Forsevere depression, which carries significantsuicide risk, the balance may bemore favorable than for anxiety and milddepression, in which suicide is lesscommon.
The American Academy of Child andAdolescent Psychiatry (AACAP) andthe American Psychiatric Association(APA) have issued guidelines for antidepressantprescribing for pediatricpatients.16 These experts recommendthat physicians closely monitor patientswho have just started taking antidepressantsin order to detect behaviorsthat could be precursors to suicidalideation or behavior. These potentialprecursors include:
Patients who experience 1 or moreof these may be at risk for worseningdepression or for suicide.
Physicians should also ask directlyabout past and present history of suicideattempts, suicidal thinking, and plansfor suicide. Ruling out bipolar disorder,including family history of bipolar disorder,is also important before antidepressanttreatment is initiated. Patientsshould be monitored once a week forthe first 4 weeks of treatment andbiweekly for the next 8 weeks. Drugswith a short half-life may cause moreadverse effects than those with a longhalf-life because of the risk of withdrawalsyndrome and worsened depressionfollowing missed doses.
Table 2 lists clinical recommendationsfor antidepressant use in childrenand adolescents, based on a variety ofsources, including guidelines and practiceparameters developed by theAACAP and the APA.16,17
Finally, it is important to rememberthat when used with appropriate cautionand monitoring, antidepressants offerhope to millions of young people sufferingfrom mental illness, who need tohave access to all available treatments.