© 2007 American Psychiatric Association
Letter to the Editor
SSRI Prescriptions and the Rate of Suicide
MARK OLFSON, M.D., M.P.H., and DAVID SHAFFER, M.D.
New York, N.Y.
To the Editor: Dr. Gibbons et al. pointed to recent trends in SSRI prescriptions and the rate of suicide in young people to suggest that the FDA warnings have contributed to the increase in the number of youth suicides from 2003 to 2004. On the basis of their analyses, the authors predicted that if the recent expansion of the FDA black box warning to young adults decreases overall SSRI prescriptions by 20%, there would be an additional 3,040 suicides in the United States over a 1-year period.
The authors reported that the national SSRI antidepressant prescription rates declined between 2004 and 2005 for all age groups, except those 60 years. In light of these declines, it is instructive to compare the national number of suicides in 2004 (1) with recently available preliminary figures for 2005 (2), overall and within the relevant age strata. Between 2004 and 2005, the total number of suicides declined from 32,439 (2004) to 31,769 (2005). More specifically, the number of suicides declined for persons ages 25 to 44 (11,712 to 11,262), ages 15 to 24 (4,316 to 4,139), and ages 5 to 14 (285 to 270) (1, 2). These declines occurred despite decreasing overall SSRI prescriptions among these age groups reported by Dr. Gibbons et al. In terms of rates per 100,000, the suicide rate for all ages declined from 11.0 (2004) to 10.7 (2005). For ages 25 to 44, the rate of suicide declined from 13.9 to 13.4, and it declined from 10.3 to 9.8 for ages 15 to 24. For ages 5 to 14, the rate of suicide remained constant at 0.7 (1, 2). The ratio of preliminary-to-final all-age suicides was 0.968 in 2002, 0.973 in 2003, and 0.976 in 2004 (2).
The focus of Dr. Gibbons et al. on SSRI prescriptions may not have captured the full range of effects of the warnings on clinical practice. For example, the warnings were associated with an increase in prescriptions of non-SSRI antidepressants to youth as physicians searched for alternative treatments (3). The effects of the warnings on the use of antipsychotic medications and other psychotropic medications remain poorly defined. Detailed longitudinal analyses of various classes of psychotropic medications and psychotherapy would enrich our understanding of the various effects of the warnings on clinical practice.
We feel that it is risky to draw conclusions from limited ecologic analyses of isolated year-to-year fluctuations in antidepressant prescriptions and suicides. One promising epidemiological approach involves examining the associations between trends in psychotropic medication use and suicide over time across a large number of small geographic regions. Until the results of more detailed analyses are known, prudence dictates deferring judgment concerning the public health effects of the FDA warnings.
Dr. Olfson has received research support from NIH, Substance Abuse and Mental Health Administration, Agency for Healthcare Research and Quality, American Foundation for Suicide Prevention, National Association for Research on Schizophrenia and Affective Disorders, New York State Office of Mental Health, Bristol-Myers Squibb, Eli Lilly, and AstraZeneca; he has also worked as a consultant for McNeil Pharmaceuticals, Pfizer, Janssen, Bristol-Myers Squibb, and Eli Lilly. Dr. Shaffer reports no competing interests.
This letter (doi: 10.1176/appi.ajp.2007.07091467) was accepted for publication in October 2007.
- Minino AM, Heron MP, Murphy SL, Kochanek KD; Division of Vital Statistics: Deaths: Final Data for 2004. Hyattsville, Md, National Center for Health Statistics, 2007
- Kung HC, Hoyert DL, Xu J, Murphy SL: Deaths: Preliminary Data for 2005. Hyattsville, Md, National Center for Health Statistics, 2005
- Nemeroff CB, Kalali A, Keller MB, Charney DS, Lenderts SE, Cascade EF, Stephenson H, Schatzberg AF: Impact of publicity concerning pediatric suicidality data on physician practice patterns in the United States. Arch Gen Psychiatry 2007; 64:466–472 [Abstract/Free Full Text]