1	Psychopharmacologic Drugs Advisory Committee and Pediatric Subcommittee of the Anti-Infective Drugs Advisory Committee February 2, 2004 Office of Drug Safety Data Resources for the Study of Suicidal Events Associated with Pediatric Use of Antidepressants Andrew D. Mosholder, M.D., M.P.H. Office of Drug Safety
2	Acknowledgements AERS Carol Pamer, R.Ph. Lynette Swartz, MBA/CIS, M.Ed. Epidemiological databases David Graham, M.D., M.P.H. Cynthia Kornegay, Ph.D. Parivash Nourjah, Ph.D. Mark Avigan, M.D. Mary Willy, Ph.D.
3	Objective To describe data resources available to ODS relevant to the issue of suicidal behaviors with pediatric use of antidepressant drugs Databases FDA postmarketing surveillance database Population-based epidemiological databases To describe context of spontaneous postmarketing reports of pediatric suicidality with newer antidepressant therapy
4	Postmarketing surveillance data from FDA’s Adverse Event Reporting System (AERS)
5	Methods AERS database search: Drugs Bupropion Citalopram Escitalopram Fluoxetine Fluvoxamine Mirtazapine Nefazodone Paroxetine Sertraline Venlafaxine Age of patient < 17 years U.S. reports only
6	Methods AERS database search: Adverse event terms (from MedDRA dictionary) Depression suicidal Gun shot wound Intentional self-injury Non-accidental overdose Overdose NOS Self injurious behavior Self injurious ideation Self-mutilation Suicidal ideation Suicide attempt Completed suicide
7	Results Total Pediatric Suicide-related events (for full marketing history of all 10 drugs) 524 total case reports 110 deaths raw counts (i.e., no hands-on review for duplicate reports of same case from more than one source)
8	Results Total raw counts of reports by drug for suicide-related terms Fluoxetine 210 Paroxetine 119 Sertraline 76 Bupropion 47 Venlafaxine 34 Citalopram 21 Fluvoxamine 10 Mirtazapine 5 Escitalopram 1 Nefazodone 1
9
10
11	Adverse events during first three years of marketing Reports for each of these 10 drugs from the first 3 years of their U.S. marketing were reviewed individually Conventional way of comparing drugs where reporting may be most similar For these drugs, time periods being compared have limited pediatric use Secular trends and FDA reporting systems changes over the time period introduce significant variability into comparison
12	Adverse events during first three years of marketing Duplicate reports eliminated Grouped into 4 suicide-related categories: Suicidal Ideation Suicide Attempt Completed Suicide Self-Mutilation
13	Results for first 3 years of marketing 94 reports reviewed 78 unduplicated reports for 9 drugs (no cases for nefazodone) Of the 78 reports Most female (77%) Most > 12 years old (95%) Most suicide attempts (86%) 7 completed suicides 6 fluoxetine, 1 paroxetine 4 males and 3 females No reports of rechallenge with same drug
14	Number of AERS reports for selected MedDRA terms, by drug, first 3 years of marketing
15	"Interpretation of AERS Reports:" Interpretation of AERS Reports: Suicidality and associated behaviors evidenced with all drugs (including paroxetine) Drugs with largest numbers of reports reflect drugs with greatest amount of pediatric use Reporting variable and influenced by media or other events Report quality, variability, and low pediatric use do not support qualitative or quantitative comparison for relative safety or risk
16	Interpretation of AERS reports, continued In general, AERS data are most useful for distinctive or rare adverse drug reactions (e.g., aplastic anemia) MOST IMPORTANT limitation for this issue: Outcome of interest (suicidality) is also an outcome of the psychiatric disorders for which these drugs are prescribed
17	Population-based epidemiological data resources
18	Databases evaluated for the potential study of pediatric suicidality and antidepressant drugs
19	Limitations of Data Sources rarity of completed suicide as an outcome difficulty in identifying individuals with outcome of completed suicide difficulty of classifying outcomes of non-fatal suicidal behavior difficulty obtaining data on drug exposure lack of suitable control/comparison groups confounding by indication privacy restrictions
20	Conclusions For the study of pediatric suicidal behavior associated with antidepressant drug treatment: Utility of available pharmacoepidemiological or postmarketing surveillance data is limited Randomized, controlled trial data should be superior to data from these sources
21