European Journal of Epidemiology 16: 789791, 2000.

2001 Kluwer Academic Publishers. Printed in the Netherlands.

Heavy coee drinking and the risk of suicide

A. Tanskanen1, J. Tuomilehto2, H. Viinamaki1, E. Vartiainen2, J. Lehtonen1 & P. Puska2
1

Department of Psychiatry, University of Kuopio, Kuopio, Finland; 2National Public Health Institute, Helsinki, Finland

Accepted in revised form 3 October 2000

Abstract. Earlier research has implicated coee

drinking as a possible protective factor for suicide.
We followed-up 43,166 subjects for the mean
14.6 years, and 213 suicides were committed. Daily
coee drinking had a J-shaped association with the

risk of suicide. Using the Cox model we controlled

for potential covariates, and found that among heavy
coee drinkers (P8 cups/day) the risk of suicide was
58% higher compared with more moderate drinkers.

Key words: Coee, Mortality, Risk, Suicide

Introduction
A signicant inverse association has been previously
reported between coee drinking and the risk of
suicide [1, 2]. This protective eect has been attributed to caeine that is a central nervous system
stimulant [3]. However, both the suicide rate and the
coee drinking level in the populations of these
studies were low. According to trade statistics, the
average annual consumption of coee per capita in
Finland was 12.4 kg (corresponding to 4 cups/day) in
1994, which is the highest amount in the world. The
suicide rate of the Finns is also high internationally,
and the rate has constantly increased during the
recent decades [4].
Therefore, we considered it important to analyze
the association between coee drinking and the risk
of suicide in the Finnish population using prospective
data from large cohorts with a long follow-up.
Material and methods
Five independent population surveys (1972, 1977,
1982, 1987, 1992) have assessed the levels of risk
factors for coronary heart disease in eastern, southwestern and southern Finland. For each survey, an
independent random sample of subjects aged 25
64 years was drawn from the National Population
Register. The participation rates varied from 69 to
96% by survey year, gender, and area. There were
20,995 male and 22,171 female subjects available for
the present analyses. Briey, the surveys included a
self-administered questionnaire (mainly questions on
socioeconomic factors, medical history, health behavior and psychosocial factors) and measurements
of height, weight and blood pressure. Details of the
survey procedures have been described earlier [5]. All
the study procedures were carried out in accordance

with the ethical standards of the Helsinki Declaration

of the World Medical Association.
The subjects were followed until 31 December
1995, or death. The mean follow-up was 14.6 (range
124) years. Information on deaths caused by suicide
was obtained from the National Death Register. The
end points were the codes E950E959 of the International Classication of Diseases (ICD). The eighth
revision of ICD was used from 1972 to 1986, and the
ninth revision was adopted from the beginning of
1987.
The amount of coee drinking was determined
with an open question: How many cups of coee do
you usually drink daily. The daily coee use was
classied into six categories (01, 23, 45, 67, 89
and 10 or more cups/day) in the rst analysis.
Subsequently, the coee drinking variable was
dichotomized to light/moderate level of drinking
(07 cups/day), and to heavy (P8 cups/day) drinking
for the multivariate analysis.
In the Cox proportional hazards regression model,
the multivariate relative risk of suicide was controlled
for potential covariates like gender, age, marital status, education, employment status, smoking, alcohol
consumption, tea drinking, and body-mass index.
The subjects with feelings of depression, anxiety,
phobia, or psychosocial stress, and with current
psychotropic medication at baseline were excluded
from the nal analysis.
Results
Eight percent of the study subjects were drinking 01,
19% 23, 30% 45, 26% 67, 9% 89 and 8% 10 or
more cups of coee daily at baseline. During the
follow-up there were 213 suicides, and the number of
suicides in the six coee drinking categories were 19,
29, 56, 48, 28 and 33, respectively.

790
Daily coee drinking had a J-shaped association
with the risk of suicide (Figure 1). The age-adjusted
risks (with 95% CI's) of suicide were 1.00, 0.67 (0.38
1.20), 0.82 (0.481.37), 0.82 (0.481.39), 1.32 (0.74
2.37) and 1.69 (0.962.98), in the six coee drinking
categories, respectively.
In the Cox proportional hazards regression model
including the potential covariates, heavy coee
drinking was signicantly associated with mortality
from suicide (Table 1). There was a 58% higher relative risk (RR: 1.58, 95% CI's: 1.022.42; p 0.039)
of suicide among heavy (P8 cups/day) coee drinkers compared with more moderate drinkers.

Discussion
Our results suggest that heavy coee drinking may be
an independent risk factor for suicide, and the association seems to be J-shaped, moderate drinkers
having the lowest risk. These results dier from previous assertions that coee drinking might be a protective factor for suicide [1, 2]. In the previous studies
the highest coee drinking categories were six [1], and
four [2] or more cups of coee per day. Consequently,
the earlier studies did not evaluate the risk of suicide
in really heavy (excessive) coee users at all. The
Finns are a population, in which both the suicide rate

Figure 1. Age-adjusted risk of suicide by daily coee drinking categories among Finnish adults (n 43,166).

Table 1. Adjusted RR of potential risk factors for suicide calculated by the Cox proportional hazards
regression model*
Variable

RR

95% CI's

p-value

Male gender
Unemployment
Single, divorced or widowed
Current smoking
Heavy coee drinking (P8 cups/day)
Heavy alcohol consumption (>120 g/week)
Low education (<7 years)
Body-mass index (kg/m2)
Age (years)
Tea drinking

4.77
3.50
1.74
1.65
1.58
1.37
1.37
1.03
1.00
0.94

2.698.44
1.597.67
1.152.63
1.092.48
1.022.42
0.842.23
0.802.35
0.971.08
0.971.02
0.601.48

0.0001
0.002
0.009
0.017
0.039
0.21
0.25
0.36
0.89
0.79

* The model included 28,040 subjects with 107 suicides. Subjects with feelings of depression, anxiety,
phobia or psychosocial stress, and those with current psychotropic medication at baseline were
excluded from the analysis.

791
and the mean level of coee drinking are clearly
higher than those in the USA, where the earlier
studies have been carried out. In addition, the concentration of coee in Finland is known to be
stronger than that usually consumed in the USA. It is
possible that these facts explain the discrepancy between our results and the previous ndings.
We were unable to calculate the total caeine intake, and this is a limitation of our study. However,
the use of soft drinks containing caeine is infrequent
among adults in Finland, so it would have not contributed much to the total caeine intake after all. All
information was gathered at only one point in time,
at baseline. So any changes in coee use, any other
risk factor, or in life circumstances generally might
have altered the coeesuicide relationship. However,
this is the limitation in most of the long-term prospective studies.
Caeine is an addictive psychoactive substance,
and heavy caeine use is inuenced by genetic factors
[6]. It increases nervousness, fear, tension, palpitations, restlessness, tremors, and may induce subjectrated anxiety and panic attacks in sensitive normal
subjects [7]. Acute intake of large amounts of caeine
may increase psychoses and hostility [8]. The plausible mechanism by which excessive coee drinking
increases the risk of suicide remains to be revealed.
The association observed may have arisen because
excessive coee drinking (i.e. caeine abuse) may be
correlated with one or more `third' factors predisposing to the mental state (e.g. personality disorder)
that raises the risk of suicide.
In conclusion, there is now new evidence that excessive coee drinking may be hazardous to mental
health. Nowadays psychiatric researchers are urged to

routinely screen for and control the use of caeine [9].

Perhaps it is time also for clinicians to do the same.
References
1. Klatsky AL, Armstrong MA, Friedman GD. Coee,
tea, and mortality. Ann Epidemiol 1993; 3: 375381.
2. Kawachi I, Willett WC, Colditz GA, Stampfer MJ,
Speizer FE. A prospective study of coee drinking and
suicide in women. Arch Intern Med 1996; 156: 521525.
3. Szekely T. Caeine as a stimulant against suicide. Arch
Intern Med 1997; 157: 243244.
4. Piribauer F, Etzersdorfer E. Recent reversal of trends in
suicide mortality in Austria and Hungary but not in
Finland. Crisis 1995; 16: 181183.
5. Vartiainen E, Puska P, Pekkanen J, Tuomilehto J,
Lonnqvist J, Ehnholm C. Serum cholesterol concentration and mortality from accidents, suicide, and other
violent causes. Br Med J 1994; 309: 445447.
6. Kendler KS, Prescott CA. Caeine intake, tolerance,
and withdrawal in women: A population-based twin
study. Am J Psychiatry 1999; 156: 223228.
7. Bruce M, Scott N, Shine P, Lader M. Anxiogenic eects
of caeine in patients with anxiety disorders. Arch Gen
Psychiatry 1992; 49: 867869.
8. Hughes JR, McHugh P, Holtzman S. Caeine and
schizophrenia. Psychiatr Serv 1998; 49: 14151417.
9. Hughes JR, Howard TS. Nicotine and caeine use as
confounds in psychiatric studies. Biol Psychiatry 1997;
42: 11841185.
Address for correspondence: Dr Antti Tanskanen, Department of Psychiatry, University of Kuopio P.O. Box 1777,
70211 Kuopio, Finland
Phone: 358-17-173546; Fax: 358-17-173549
E-mail: antti.tanskanen@kuh.