Modeling of the temporal patterns of fluoxetine prescriptions and suicide rates in the United States

BACKGROUND To study the potential association of antidepressant use and suicide at a population level, we analyzed the associations between suicide rates and dispensing of the prototypic SSRI antidepressant fluoxetine in the United States during the period 1960-2002. METHODS AND FINDINGS Sources of data included Centers of Disease Control and US Census Bureau age-adjusted suicide rates since 1960 and numbers of fluoxetine sales in the US, since its introduction in 1988. We conducted statistical analysis of age-adjusted population data and prescription numbers. Suicide rates fluctuated between 12.2 and 13.7 per 100,000 for the entire population from the early 1960s until 1988. Since then, suicide rates have gradually declined, with the lowest value of 10.4 per 100,000 in 2000. This steady decline is significantly associated with increased numbers of fluoxetine prescriptions dispensed from 2,469,000 in 1988 to 33,320,000 in 2002 (r(s) = -0.92; p < 0.001). Mathematical modeling of what suicide rates would have been during the 1988-2002 period based on pre-1988 data indicates that since the introduction of fluoxetine in 1988 through 2002 there has been a cumulative decrease in expected suicide mortality of 33,600 individuals (posterior median, 95% Bayesian credible interval 22,400-45,000). CONCLUSIONS The introduction of SSRIs in 1988 has been temporally associated with a substantial reduction in the number of suicides. This effect may have been more apparent in the female population, whom we postulate might have particularly benefited from SSRI treatment. While these types of data cannot lead to conclusions on causality, we suggest here that in the context of untreated depression being the major cause of suicide, antidepressant treatment could have had a contributory role in the reduction of suicide rates in the period 1988-2002.The authors were supported by NIH grants GM61394, RR017365, MH062777, RR000865, K30HL04526, K12RR17611, RR16996, HG002500, and DK063240, and by an award from the Dana Foundation