Is Antidepressant Prescribing Associated With Suicide Rates?

May 1, 2004

Since the introduction of the SSRIs in the early 1990s, the rate of antidepressant prescribing has increased dramatically. This look at five national data sets concludes that greater recognition of depression and greater rates of treatment with medication and psychosocial interventions has made a significant contribution to reducing suicide rates.

Since the introduction of the selective serotonin reuptake inhibitors and other new antidepressants in the early 1990s, there has been a major shift in prescribing patterns for this class of medications. Rather than the newer agents merely replacing the former market-leading class (the tricyclic antidepressants)--as has happened for other therapeutic groups in medicine such as the antihypertensives--there has been a true increase in usage of antidepressants.

For example, in Australia during the period from 1990 to 1998, there was a 50% increase in prescriptions for antidepressants and a threefold increase in the number of defined daily doses of antidepressants per unit of population (McManus et al., 2000). (The defined daily dose is the World Health Organization figure for the average therapeutic dose of a compound. For example, the defined daily dose for fluoxetine [Prozac] is 20 mg. The finding of 35.7 defined daily doses of antidepressants per 1,000 population in 1998 by McManus et al. [2000] equates to a mean of 35.7 individuals per thousand population receiving a therapeutic dose of any antidepressant.) Using commercial data, McManus et al. (2000) also reported similar increases in most developed nations, including the United States and Europe. Moreover, in Australia, this appeared to be mainly a phenomenon in primary care, with 86% of antidepressant prescribing taking place in general practice (McManus et al., 2003). More recent data indicate that this increased prescribing is continuing (Mant et al., in press), although there is some suggestion that the rate of incline may be slowing.

In the United States, a series of reports confirmed significant increases in outpatient diagnosis and antidepressant treatment of depression during the 1990s (e.g., Olfson et al., 2002; Pincus et al., 1998). Middleton et al. (2001) found similar changes in the United Kingdom, reporting a more than twofold increase in antidepressant prescriptions for the period 1975 to 1998. They found increases in prescribing in all age and gender groups, but most noticeably a threefold increase in older age groups. While the absolute volume of antidepressant prescribing remained greater for females than males over that period of time, the increases were more marked in males, with the exception of the 12- to 19-year-old age group.

Have there been, however, demonstrable beneficial outcomes of such increased prescribing patterns? A number of large-scale national epidemiological studies have been consistent in identifying substantial rates of untreated or poorly treated depression (Kessler et al., 2003), indicating the likely profound benefits to entire communities and populations of improved rates of identification and treatment of depression.

Yet, despite this, there has been considerable lay and professional disquiet at the increased usage of antidepressants. Furthermore, until recently, there has been little evidence that this increased antidepressant prescribing has had any public health impact. In 2003, the eminent European mood disorder researcher Herman van Praag lamented what he has termed the "stubborn behaviour" of the failure of antidepressants to reduce suicide rates, arguing that the majority of evidence does not support any reduction in such outcomes related to the increased volume of prescribing.

Van Praag highlighted two major issues. First, there have been few countries reporting sustained reductions in suicide rates, despite the substantial increase in antidepressant prescribing. Second, randomized, controlled trials of antidepressants have failed to demonstrate any effect on suicide. In a most telling report, Khan et al. (2003) analyzed U.S. Food and Drug Administration summary reports of controlled clinical trials for nine recently marketed antidepressants. Similar suicide rates were seen in those randomly assigned to each SSRI, comparator antidepressant or placebo, failing to provide any evidence of a suicide-reducing effect of antidepressants. There are, however, limitations as to how much one can infer from such trials. Suicide is a rare event, and even in the FDA trials, which comprised over 48,000 patients with depression, only 77 committed suicide, making comparisons between treatment groups difficult. Moreover, such trials exclude entry to those with prior significant suicidal risk.

Despite such negative findings, a potential beneficial effect of antidepressants on suicide rates has become apparent in a recent series of national population-based reports. The first studies came from countries that had observed overall reductions in their suicide rates. Isacsson (2000) and Carlsten et al. (2001) both investigated the Swedish experience, where there had been a gradual reduction in suicide rates over the period from 1977 to 1997. They found that the decline in suicides accelerated after 1990 when the SSRIs were introduced. The rate of suicides in the 1990s was significantly inversely related to the rate of antidepressant prescribing in most age and gender groups. Similarly, in Hungary, rates of suicide declined in parallel with a rapid growth of antidepressant usage, despite steep increases in unemployment and per capita alcohol consumption (Rihmer, 2001, as cited in Hall et al., 2003). The same phenomenon was, however, not observed in all European countries, with no such association being seen in Italy (Barbui et al., 1999).

It was against this background that we decided to investigate whether there was any substantive improvement in suicide rates or other public health benefit from the dramatic increase in antidepressant prescribing in Australia. We examined the association between changes in antidepressant prescribing in Australia for the period 1991 to 2000 (Hall et al., 2003).

One of the complexities that faced us, however, was that the total suicide rate for Australian men and women did not change between 1991 and 2000, because marked decreases in suicide rates in older men and women were offset by increases in younger adults, especially young men. A similar phenomenon has been observed in the United Kingdom, where suicide rates doubled in males younger than 45 between 1950 and 1998, but rates declined in older males and females of all ages (Gunnell et al., 2003). Because of this phenomenon, we analyzed differences in suicide trends between men and women in different age groups to assess whether age and gender rates in suicide were related to differences between these groups in exposure to antidepressant medication (Hall et al., 2003).

The methodology that we used was a quasi-experimental approach to analyze for associations using a number of different prospectively collected Australian national data sets. First, to determine suicide rates in different age and gender groups, we were provided with data from the Australian Bureau of Statistics on all cases of suicide for each year for 1986 to 1990 (the baseline pre-SSRI period) and 1996 to 2000 (the post-introduction of SSRI period). Annual mortality for men and women in eight age groups (Table) was calculated. Second, to determine antidepressant use in those two time periods, we combined data on total sales of antidepressants (obtained from commercial sources) with estimates from large national general practice studies of the proportion of antidepressants that were prescribed to men and women in these age groups. This enabled us to determine defined daily doses/1,000 population for each of those time periods. We then used Spearman rank correlations to assess associations between trends in rates of suicide with trends in antidepressant prescribing for each gender. The trends in antidepressant prescribing that we examined were total exposure over the period of time and the rate of change in prescribing over the study period.

We found strong evidence of a beneficial impact of antidepressant prescribing on suicide rates. Among both men and women, the largest declines in suicide occurred in the age groups with the highest exposure to antidepressants across the study period (males, r=-0.91; females, r=-0.76; both significant) (Figure 1). Furthermore, as detailed in Figure 2, there was also a significant inverse correlation between change in the defined daily dose/1,000 for women (r=-0.74), with a trend toward significance for males (r=-0.62).

It was not possible to explain our findings in terms of concurrent changes in social factors that are recognized determinants of suicide rates. For example, per capita alcohol consumption remained steady over most of the study period, and unemployment rates in older males increased over that time.

We felt that it was unlikely that the antidepressants alone were responsible for the reduction in suicide rates and would rather view antidepressant prescribing in primary care as a proxy for exposure to psychosocial and pharmacological interventions, as data from Australian general practice surveys have indicated that general practitioners rarely provide antidepressants without some assessment and management of relevant psychosocial issues.

The validity of our findings has been supported by two other studies published in 2003. In the United States, Olfson et al. (2003) evaluated the relationship between geographical regional changes in antidepressant prescribing and suicide in adolescents from 1990 to 2000. A significant negative relationship (after adjusting for potential confounds such as gender, age, income and race) was found between regional change in antidepressant medication treatment and suicide during the study period. They calculated that a 1% increase in adolescent use of antidepressants was associated with a decrease of 0.23 suicides per 100,000 adolescents per year. The clearest benefits were found in males, youths aged 15 to 19 and those with lower family incomes. In the United Kingdom, Gunnell et al. (2003), examining the period 1950 to 1998, found that the dramatic reductions in suicide rates in older people were associated with increases in gross domestic product, the size of the female work force, marriage and the increased prescribing of antidepressants. This report highlighted that the population trends in suicide appeared to be associated with a range of social and health-related factors, with antidepressants comprising one of the latter.

It is therefore apparent from these studies of five different national data sets that the recent phenomenon of a substantial increase in the recognition of depression and greater rates of treatment with antidepressant medications and psychosocial interventions appears to be one of the significant contributants to reduced suicide rates in either total populations, older people or adolescents. These findings give encouragement that improved therapies for depression and improved professional and community awareness of depression are beginning to pay dividends in terms of reduced suicide rates. However, suicide rates remain unacceptably high in all countries. We have a long way to go.




Barbui C, Campomori A, D'Avanzo B et al. (1999), Antidepressant drug use in Italy since the introduction of SSRIs: national trends, regional differences and impact on suicide rates. Soc Psychiatry Psychiatr Epidemiol 34(3):152-156.


Carlsten A, Waern M, Ekedahl A, Ranstam J (2001), Antidepressant medication and suicide in Sweden. Pharmacoepidemiol Drug Saf 10(6):525-530.


Gunnell D, Middleton N, Whitley E et al. (2003), Why are suicide rates rising in young men but falling in the elderly?--A time-series analysis of trends in England and Wales 1950-1998. Soc Sci Med 57(4):595-611.


Hall WD, Mant A, Mitchell PB et al. (2003), Association between antidepressant prescribing and suicide in Australia, 1991-2000: trend analysis. BMJ 326(7397):1008 [see comments].


Isacsson G (2000), Suicide prevention--a medical breakthrough? Acta Psychiatr Scand 102(2):113-117 [see comment].


Kessler RC, Berglund P, Demler O et al. (2003), The epidemiology of major depressive disorder: results from the National Comorbidity Survey Replication (NCS-R). JAMA 289(23):3095-3105 [see comments].


Khan A, Khan S, Kolts R, Brown WA (2003), Suicide rates in clinical trials of SSRIs, other antidepressants, and placebo: analysis of FDA reports. Am J Psychiatry 160(4):790-792 [see comment].


Mant A, Rendle VA, Hall WD et al. (in press), Making new choices about antidepressants in Australia: the long view 1975-2002. Med J Aust.


McManus P, Mant A, Mitchell P et al. (2003), Use of antidepressants by general practitioners and psychiatrists in Australia. Aust N Z J Psychiatry 37(2):184-189.


McManus P, Mant A, Mitchell PB et al. (2000), Recent trends in the use of antidepressants in Australia, 1990-1998. Med J Aust 173(9):458-461 [see comments].


Middleton N, Gunnell D, Whitley E et al. (2001), Secular trends in antidepressant prescribing in the UK, 1975-1998. J Public Health Med 23(4):262-267.


Olfson M, Marcus SC, Druss B et al. (2002), National trends in the outpatient treatment of depression. JAMA 287(2):203-209 [see comments].


Olfson M, Shaffer D, Marcus SC et al. (2003), Relationship between antidepressant medication treatment and suicide in adolescents. Arch Gen Psychiatry 60(10):978-982.


Pincus HA, Tanielian TL, Marcus SC et al. (1998), Prescribing trends in psychotropic medications: primary care, psychiatry, and other medical specialities. JAMA 279(7):526-531 [see comments].


van Praag HM (2003), A stubborn behaviour: the failure of antidepressants to reduce suicide rates. World J Biol Psychiatry 4(4):184-191.