Oct 27, 2000 - suicide attempters died during the average 5.3-year follow-up of the index attempt. Deaths from suicide accounted for 37% of all excess deaths ...
Soc Psychiatry Psychiatr Epidemiol (2001) 36: 29±35
Ó Steinkopff-Verlag 2001
ORIGINAL PAPER A. Ostamo á J. LoÈnnqvist
Excess mortality of suicide attempters
Accepted: 27 October 2000
j Abstract Background: Excessive mortality of suicide attempters has emerged from many follow-up studies. Completed suicide is the main cause of excess deaths, but the increased risk of deaths from other unnatural and natural causes is also of major public health concern. We lack follow-up studies of the different causes of death in cohorts of suicide attempters. The present study aimed to determine the mortality by suicide and other causes of death and to investigate risk factors. Methods: This mean 5.3year follow-up study was based on an unselected cohort of suicide attempts by both violent and nonviolent methods, treated in hospitals in a well-de®ned urban catchment area in Helsinki. In total, 2782 patients aged 15 years and over admitted to the emergency rooms after suicide attempt between 1989 and 1996 were included in the follow-up analysis. Standardised mortality ratios (SMR) for suicide, disease, accident, homicide, and undetermined death were calculated. Results: Mortality from all causes was 15 times higher than that expected among men and nine times higher in women. SMRs in men were 5402 (95% CI 4339±6412) for suicide, 2480 (95% CI 925± 4835) for homicide, and 11,139 (95% CI 6884±16,680) for undetermined cause, and for women 7682 (95% CI 5423±9585), 3763 (95% CI 52±5880) and 15,681 (95% CI 6894-22,294), respectively. Fifteen percent of all suicide attempters died during the average 5.3-year follow-up of the index attempt. Deaths from suicide accounted for 37% of all excess deaths in men and 44% in women. The mortality ratio was highest
Introduction Excessive mortality of suicide attempters, especially by suicide, has been found in many follow-up studies (Hawton and Fagg 1988; Harris and Barraclough 1998). It is vital for suicide prevention to obtain better knowledge for assessing and treating suicidal patients, and thus the aim of most follow-up studies on attempted suicide is to improve the prediction of later suicide. However, there also seems to be a substantial increased risk of premature death among suicide attempters by other unnatural and natural causes (Pederson et al. 1975; Hawton and Fagg 1988). Follow-up studies are often based on selected populations of attempted suicide. Some have concentrated on serious suicide attempts (Rosen 1976), while others have included only attempts by a certain method, mostly self-poisoning (Ekeberg et al. 1994; Rygnestad 1997). Some studies have involved only psychiatrically hospitalised patients (Cullberg et al. 1988), while others have con®ned themselves to patients from a single general hospital (LoÈnnqvist
SPPE 427
A. Ostamo (&) á J. LoÈnnqvist National Public Health Institute, Department of Mental Health and Alcohol Research, Mannerheimintie 166, FIN-00300 Helsinki, Finland e-mail: aini.ostamo@ktl.® Tel.: +358-9-47448389; Fax: +358-9-47448478
during the 1st follow-up year. The total number of lost years of life among the 413 suicide attempters who died during follow-up was 13,883. The risk factors for all causes of death were male sex, single, retirement, drug overdose as a method, an index attempt not involving alcohol, and a repeated attempt. Conclusion: A suicide attempt indicates a severe risk of premature death, and suicide is the main cause of excess deaths. However, it appears that concentrating ef®cient treatment only on the most suicidal patients could prevent no more than two of ®ve premature deaths. More effort is therefore needed to prevent the excess mortality of suicide attempters by also addressing causes of death other than suicide.
30
and Tolppanen 1985), or to patients only from general hospitals while excluding those treated elsewhere (Hawton and Fagg 1988). Unfortunately, we lack follow-up studies on the entire cause of death spectrum in unselected cohorts of suicide attempters. The present study aimed to determine the overall mortality and causes of death, and their risk factors, in an unselected cohort of suicide attempters in Helsinki, Finland, using an average of over 5 year's follow-up.
Subjects and methods We studied all attempted suicide admissions (n 3796) of Helsinki residents aged 15 years or more in study periods covering 1989± 1996. The data were gathered for the whole calendar year of 1989, for 7 months in 1990 (1 January±31 July) and for 2-month periods (15 April±14 June) in the years 1991±1996. The treatment of suicide attempts in Helsinki (520,000 inhabitants) begins by all attempts being referred to the emergency rooms of three hospitals under
Table 1 Characteristics of the index attempt by status at the end of followup
Helsinki University Central Hospital and two municipal hospitals. General practitioners, psychiatric hospitals and other health care facilities also refer all suicide attempts to these hospitals. The data gathering took place in all ®ve hospitals treating suicide attempters in the catchment area. Thus the data were unselected and included all attempts referred to health care in Helsinki, psychiatric and nonpsychiatric patients, and suicide attempts by both violent and non-violent methods. The index attempt was the individual's ®rst admission to the hospital due to a suicide attempt during 1989± 1996, and this sets the time when the person entered the data. The mortality of the study cohort was updated from Statistics Finland. All deaths until 30 June 1997 among persons included in the study were identi®ed via personal identi®cation numbers. The of®cial death certi®cates were used to extract time and cause of death. The mean follow-up time was 5.3 years (SD 2.7, minimum 1.0 year, maximum 8.4 years). Of the subjects in the original monitoring data, 3.2% (n 93) could not be identi®ed by Statistics Finland, and 16 persons had died as a consequence of the index attempt. They were excluded from the survival analysis. Thus, of the original 2891 patients in the monitoring data, 2782 persons ± 1402 men and 1380 women ± were included in the follow-up analysis. During 1989±1997 Helsinki was one of 16 centres in the WHO/ Euro Multicentre Study on Parasuicide started in 1989 (Platt et al. 1992; Schmidtke et al. 1996; World Health Organization 1999). The
Dead (n = 413) % Sex Male 69.0 Age 15±24 9.0 25±34 24.0 35±44 28.1 45±54 14.8 55±64 9.0 65± 15.2 Marital status Single 52.5 Married 32.9 No longer married 7.1 Not known 7.5 Educational level Lower education or unknown 69.9 Upper secondary education 22.0 Higher education 8.1 Employment status Employed 30.7 Unemployed 10.9 Pensioner 33.7 Other 9.0 Not known 15.7 Method of index attempt Drug overdose 90.1 Other method 9.9 Alcohol involved in index attempt Yes 54.0 No 26.6 Not known 19.4 Number of attempts during study periods 1 82.6 2 or more 17.4
Alive (n = 2369) %
Total (n = 2782) %
Significance v2 test
47.1
50.4
0.001
18.8 33.5 27.0 12.9 5.0 2.8
17.3 32.1 27.2 13.2 5.6 4.7
0.001
56.2 29.1 3.1 11.6
55.7 29.7 3.7 11.0
0.001
68.9 24.6 6.5
69.0 24.2 6.8
0.322
42.8 12.0 14.7 14.3 16.2
41.1 11.8 17.5 13.5 16.1
0.001
85.1 14.9
85.8 14.2
0.007
60.5 21.6 17.9
59.5 22.4 18.1
0.032
90.2 9.8
89.1 10.9
0.001
31 background variables of the year cohorts from 1989 to 1992 in Helsinki have been described previously as a part of the monitoring material of the WHO study without any follow-up data (Schmidtke et al. 1996). Chi-square tests were used to test for statistical differences between deceased and living groups. Standardised mortality ratios (SMRs) were calculated to compare the mortality of suicide attempters with that of the general population. The expected values for all six broad categories of mortality (overall mortality, as well as deaths from disease, suicide, accident, violent and undetermined causes) were calculated on the basis of the average age SMRs per year by age and sex in the male and female general population of Helsinki in 1989±1996. The expected values were multiplied by 5.3, i.e. the average follow-up time. SMRs were calculated by dividing the observed by expected number of deaths and multiplying by 100. Tests of the statistical signi®cance of the SMR were based on the Poisson distribution (two-tailed), using 95% con®dence intervals. Survival analysis and the Cox proportional hazards regression model were used to analyse survival time and risk factors for each broad cause of death group. The follow-up time started when the person made his/her index attempt and ended either at death or the end of the follow-up period, i.e. 30 June 1997. The event of interest was death.
Results By the end of the follow-up, 14.8% of suicide attempters (n 413) 20.3% of men (n 285) and 9.3% of women (n 128) had died (Table 1). Twothirds were men, and the deceased overall were older (mean age 44.8 years in men and 50.2 in women) than the living (mean age 40.5 years in men and 42.4 in women). They were also signi®cantly more often married or no longer married. There were no signi®cant differences between the deceased and those alive by education level, but the dead were signi®cantly more often retired. The deceased had used drug overdose signi®cantly more often as their index attempt method than those alive at the end of follow-up. They had more seldom consumed alcohol in connection with the index attempt than others, and had repeated attempts signi®cantly more often during the follow-up period. Mortality among the suicide attempters from all causes was 15 times that expected in men and nine times that expected in women (Table 2). The mortality risk for suicide was 54 times that expected in men and 77 times that expected in women. The risk of other violent deaths was also relatively higher among women than men: 38 versus 25 times higher than expected. SMRs were particularly high for undetermined deaths. SMRs for the suicide attempters showed a greater risk for women than men for suicide, homicide and undetermined causes of death. Mortality from natural causes (diseases) was six times that expected in men and four times that expected in women (Table 2). The most common causes of death in this group were diseases of the circulatory system, pneumonia or other lung diseases, and cancer. The underlying or contributory causes of
Table 2 Observed and expected deaths and standardised mortality ratios (SMRs) of suicide attempters (n = 2928) by cause of death during on average, 5.3 years' follow-up Cause of death All causes Men Women Suicide Men Women Disease Men Women Accident Men Women Homicide Men Women Undetermined Men Women
Observed no. of deaths
Expected no. of deaths
SMR
95% CI
285 128
19.5 13.6
1462* 938*
1296±1642 785±1119
106 51
2.0 0.7
5402* 7682*
4339±6412 5423±9585
88 46
14.9 12.1
592* 379*
474±728 278±507
62 15
2.2 0.7
2847* 2120*
1590±3614 1196±3543
7 3
0.3 0.1
2480* 3763
925±4835 52±5880
22 13
0.2 0.1
11139* 15681*
6884±16680 6894±22294
* Statistically significant
death were often explicitly attributed to alcohol: alcohol dependence, alcoholic cardiomyopathy, alcoholic diseases of the liver and pancreas, alcoholic epilepsy and diabetes. Thirty-six percent of all death certi®cates mentioned alcohol dependence or abuse. Suicide was the most common cause of death among both sexes. Male attempters died twice (7.5%) as often by subsequent suicide as females (3.7%). Among the deceased from natural causes, women (36%, n 46) had died slightly more often than men (31%, n 88), whereas fatal accidents among men (22%, n 62) were twice as common (women 12%, n 15). The cause of death was suicide among two®fths of both men (37%, n 106) and women (40%, n 51). Seven men and three women had died by homicide. The cause was unclear in 8% (n 22) of male and 10% (n 13) of female deaths. All-cause mortality was highest during the 1st follow-up year, at 4.8% ± 6.6% among men (Fig. 1) and 3.0% among women (Fig. 2). Throughout the follow-up, the risk of all-cause mortality among females was lower. Suicide mortality was particularly high in the 1st follow-up year, at 2.4% ± 3.6% among men and 1.2% among women. Survival analysis showed that men had a signi®cantly higher risk of suicide (log rank 20.762, df 1, p 0.0001), disease (log rank 14.9181, df 1, p 0.0001), accident (log rank 30.3178, df 1, p 0.0001), and overall mortality (log rank 64.9674, df 1, p 0.0001). Male sex was the main risk factor for premature death in the ®nal Cox models, both for all-cause mortality and for each broad cause of death group
32 Fig. 1 Cumulative survival by cause of death among male suicide attempters during an 8.4-year follow-up
Follow-up year
Fig. 2 Cumulative survival by cause of death among female suicide attempters during an 8.4-year follow-up
Follow-up year
separately, undetermined deaths being the exception (Table 3). In all-cause mortality, the risk factors were male sex, single, retirement, drug overdose as the attempted suicide method, and repeated attempts. The risk increased with age. If suicide attempters, especially men, did not have alcohol in the blood when referred to hospital, and if they had made repeated attempts during the monitoring periods, the relative risk for suicide was most pronounced. Age was not a signi®cant predictor of suicide. Male sex also carried greater relative risk of death from natural causes, and this increased with age. Being a man with a low level of education and repeated suicide attempts predicted accidental death. As for undetermined deaths, repeated attempts was the only signi®cant predictor, with a relative risk of 2.7. The crude estimation of lost years of life in this attempted suicide population, compared to the mean expectation of life in the general Finnish population in 1991±1995 (Statistics Finland 1999) and assuming the cause of death to be independent of all other
causes of death, averaged 33 years per person among men and 34 years among women. The total number of lost years of life among the 413 suicide attempters who died during the follow-up was 13,883.
Discussion We found a considerable excess mortality among suicide attempters from both natural and unnatural causes, especially the latter. In this study 15% of the suicide attempters, mean age 36 years in men and 38 years in women, died during the 5.3 years follow-up after the index attempt. Deaths from suicide accounted for 37% of excess mortality in men and 44% in women. The male attempters had greater mortality for all main causes of death. Compared to the general population, the SMRs showed a greater risk for women than men of dying from suicide, homicide and unclear causes of death. Although the data gathering in 1990±1996 was based only on samples, the stable admission ®gures
33 Table 3 Risk factors by cause of death: Cox proportional hazard regression model
Risk factor All deaths Age Male sex Repetitions Single Drug overdose Retired Suicide Repetitions Male sex No alcohol Disease Age Male sex Accident Male sex Age Low education Repetitions Undetermined Repetitions
Regression coefficient
Standard error
Wald statistic
df
Significance
Relative risk
Confidence interval 95%
0.042 0.999 0.618 0.342 0.474 0.322
0.004 0.108 0.131 0.114 0.166 0.133
93.270 84.932 22.085 8.992 8.167 5.903
1 1 1 1 1 1
0.0001 0.0001 0.0001 0.0027 0.0043 0.0151
1.043 2.715 1.855 1.408 1.606 1.381
1.03±1.05 2.19±3.36 1.43±2.40 1.13±1.76 1.16±2.22 1.06±1.79
0.917 0.789 0.382
0.187 0.172 0.182
24.122 20.983 4.419
1 1 1
0.0001 0.0001 0.0355
2.501 2.201 1.465
1.74±3.61 1.57±3.08 1.03±2.09
0.091 1.205
0.005 0.190
311.351 40.417
1 1
0.0001 0.0001
1.096 3.338
1.08±1.11 2.30±4.84
1.499 0.028 0.847 0.621
0.290 0.008 0.305 0.298
26.731 12.164 7.717 4.341
1 1 1 1
0.0001 0.0005 0.0055 0.0372
4.477 1.029 2.333 1.862
2.54±7.90 1.01±1.04 1.28±4.24 1.04±3.34
0.981
0.388
6.399
1
0.0114
2.668
1.25±5.71
among men and women and the stable age distributions in study cohorts are evidence of the reliability of the data (Ostamo et al. 2000). The 2-month periods chosen seem long enough to have minimised temporary ¯uctuations. According to Finnish law, a medicolegal investigation of the cause and manner of death should be carried out when death is unnatural, or suspected to be so. Medicolegal autopsies are carried out whenever possible by a small group of well-trained forensic pathologists. The autopsy rate in this study was 81%, and a medicolegal investigation was carried out by 32 forensic pathologists, seven of whom carried out 67% of all investigations. The death certi®cates are ®rst checked by legal medical of®cers at the provincial level, and then collected and checked also by Statistics Finland. All-cause mortality includes all six main categories (disease, accident, suicide, homicide, undetermined, and war when appropriate) used in Finnish death certi®cates. The overall autopsy rate in Finland is relatively high: 31% in 1995 (Central Statistical Of®ce of Finland 1996). The corresponding ®gures for Denmark, Sweden and Norway in the mid1980s were similar: 36%, 30% and 20%, respectively (Hesso 1987). Thus, determination of the cause and manner of death in Finland can be considered reliable. In a recent Finnish study, the mortality from undetermined deaths was 10% of the total suicide rate, and many of the undetermined deaths resembled È hberg and LoÈnnqvist 1998). Poisoning was suicides (O the most common cause of undetermined deaths, and 38% of those whose death had been recorded this way
had displayed some direct evidence of suicidal intent shortly before death. In our data undetermined deaths comprised 8% of all deaths, and the ratio of suicide to undetermined deaths was 4.5. It is probable that our sample also included about another 10% of suicides, which had to be classi®ed as undetermined deaths due to lack of information. Overall mortality was greatest during the 1st follow-up year, when every second suicide was committed, after which the risk declined markedly. This accords with earlier studies: both repeaters and ®rst evers have the greatest risk of suicide during the 1st year after the index attempt (Hawton and Fagg 1988; LoÈnnqvist and Ostamo 1991). It is thus extremely important for more attention to be focused on treatment during the 1st year after any suicide attempt. Thereafter, the rate of completed suicide following the index attempt remained at 1±2% annually, which has been found to be the case over the 10 years after a suicide attempt in other studies (Cullberg et al. 1988; Nordentoft et al. 1993; Tejedor et al. 1999). A repeated attempt was a signi®cant risk factor for all the causes of death except natural causes. Higher SMRs were also associated with repeated attempts in a study by Harris and Barraclough (1998), and the number of previous attempts has been found to decrease by survival time (Tejedor et al. 1999). Drug overdose as the attempted suicide method was a signi®cant risk factor for death from all causes in our study. Compared to follow-up studies on attempted suicide by any method (Ettlinger 1975; Harris and Barraclough 1998), many of
34
those limited to self-poisoning (Nordentoft et al. 1993; Ekeberg et al. 1994; Harris and Barraclough 1998) have shown higher SMRs for other causes of death than suicide. Completed suicide during the follow-up was associated with male sex, index attempt not involving alcohol, and with repeated suicide attempts. Alcohol was closely connected with suicide attempts, but those not involving alcohol have been found to associate with lower somatic seriousness (Suokas and LoÈnnqvist 1991), although the use of alcohol as part of the attempt itself adds to the danger of fatal overdose (Hawton et al. 1989). According to Harris and Barraclough (1998) all mental disorders carry an increased risk of premature death. It is evident that much of the excess death of suicide attempters is associated with high rates of concurrent mental and physical illness. In a Finnish study on suicide attempters, 82% suffered from comorbid mental disorders (Suominen et al. 1996). In many studies, mood disorders, alcoholism, schizophrenia and personality disorders have been among the most frequent primary diagnoses of suicide attempters, while panic and other anxiety disorders, drug addiction and adjustment disorders have also been associated with attempted suicide and suicidal ideation (Hawton and Fagg 1988; Hawton et al. 1989; Weissman et al. 1989; Petronis et al. 1990; Nordentoft et al. 1993; Rudd et al. 1993; Bronish and Wittchen 1994; Cox et al. 1994; Cornelius et al. 1995). In addition to suicide, premature death may result from accident proneness, coexisting substance misuse, deprivation, side effects of various medications and a high incidence of concurrent physical illness (Corten et al. 1991; Mehtonen et al. 1991; Lereya et al. 1995). In our study, the deceased tended to suffer from mental and/or physical illnesses particularly associated with alcohol consumption. Sudden death from natural causes has been reported among psychiatric patients, and may re¯ect a general neglect of health, an increased rate of damaging behaviours such as smoking and poor diet, and decreased access to health services (Corten et al. 1991; Berren et al. 1994). Research clearly reveals excess suicide risk in lower social classes and reduced suicide and attempted suicide risk in higher social classes (Platt and Hawton 2000). Studies of socio-economic differentials in health routinely show that the lower the socioeconomic status, the poorer the health (Fox 1989; Illsley and Svensson 1990). The socio-economic link with morbidity (Blaxter 1990; Lundberg 1990; Dahl 1991; Mackenbach et al. 1997) is broadly similar to that with mortality (Valkonen 1992; Fox et al. 1985). Although in this study we were unable to control for the socio-economic status of suicide attempters, other ®ndings based on the same data showed that suicidal
persons with low education were at greater risk of becoming unemployed or dropping out of the labour market (Ostamo et al. 2000), and that socio-economic disadvantage within the urban subareas of Helsinki correlated with higher attempted suicide rates (Ostamo, unpublished results). The socio-economic status of suicide attempters may be one reason for the elevated SMRs for deaths from other causes than suicide. The high mortality of suicide attempters in our study is in accordance with previous follow-up studies (Ettlinger 1975; Harris and Barraclough 1998). However, mortality risk both from all causes, and from unnatural and natural causes of death, was considerably higher in our study than in previous follow-up studies. Our cohort represents attempted suicides in a well-de®ned urban catchment area. Completed data on mortality rates for the general population of 15 years and over by sex and age and for each cause of death during the follow-up were used in this study. However, attempts not referred to health care were missing. This means that our sample represents the more severe suicide attempts in the spectrum of all attempts, which probably accounts for the high mortality rates in our study. Suicide attempts lead to a severe risk of subsequent premature death and cause many lost years of life, at least during the ®rst years of follow-up. Suicide is the main cause of excess deaths among attempters. However, it appears that concentrating ef®cient treatment only on the most suicidal patients could prevent no more than two of ®ve premature deaths. More effort is needed to prevent the excess mortality of suicide attempters from other causes of death, too, not only by training health care professionals in the recognition of suicidality and depression, but also by focusing more on developing feasible treatment strategies for psychiatric disorders, physical illnesses and substance abuse among suicide attempters. Treatment of such patients is demanding, because a comprehensive analysis of all possible risk factors of mortality is needed ± not only those directly connected with psychiatric disorders or physical illnesses, but also social and cultural factors. The urgency is ampli®ed by the fact that the speci®c psychiatric needs of attempted suicide patients are poorly met, especially concerning the treatment of depression and substance use problems (Suominen et al. 1999). The role of the health care system here is to provide quali®ed professional psychiatric treatment for all persons referred following a suicide attempt. j Acknowledgements We thank Eero Lahelma for helpful suggestions on earlier drafts and VaÈestoÈrekisterikeskus for permission to use the mortality data (VRK-690/40/97). The study was supported by the Academy of Finland (grants 2041091 and 34432).
35
References Berren MR, Hill KR, Merikle E, Gonzalez N, Santiago J (1994) Serious mental illness and mortality rates. Hosp Community Psychiatry 45: 604±605 Blaxter M (1990) Health and lifestyles. Tavistock/Routledge, London Bronish T, Wittchen H (1994) Suicidal ideation and suicide attempts: comorbidity with depression, anxiety disorders, and substance abuse disorder. Eur Arch Psychiatry Clin Neurosci 244: 93±98 Central Statistical Of®ce of Finland (1996) Statistical Yearbook of Finland 1966. Gummerus, JyvaÈskylaÈ Cornelius JR, Salloum IM, Mezzich J (1995) Disproportionate suicidality in patients with comorbid major depression and alcoholism. Am J Psychiatry 152: 358±364 Corten P, Ribourdouille M, Dramaix M (1991) Premature death among outpatients at a community mental health center. Hosp Community Psychiatry 42: 1248±1251 Cox BJ, Direnfeld DM, Swinson RP, Norton GR (1994) Suicidal ideation and suicide attempts in panic disorder and social phobia. Am J Psychiatry 151: 882±887 Cullberg J, Wasserman D, Stefansson C-G (1988) Who commits suicide after a suicide attempt? An 8 to 10 year follow-up in a suburban catchment area. Acta Psychiatr Scand 77: 598±603 Dahl E (1991) Inequality in health and the class position of women ± the Norwegian experience. Sociol Health Ill 13: 492±505 Ekeberg O, Ellingsen O, Jacobsen D (1994) Mortality and causes of death in a 10-year follow-up of patients treated for selfpoisoning in Oslo. Suicide Life Threat Behav 24: 398±405 Ettlinger R (1975) Evaluation of suicide prevention after attempted suicide. Acta Psychiatr Scand Suppl 260 Fox J (1989) Health inequalities in European countries. Gower, Aldershot Fox J, Goldblatt P, Jones D (1985) Social class mortality differentials: artefact, selection or life circumstances? J Epidemiol Commun H 39: 1±18 Harris EC, Barraclough B (1998) Excess mortality of mental disorder. Br J Psychiatry 173: 11±53 Hawton K, Fagg J (1988) Suicide, and other causes of death, following attempted suicide. Br J Psychiatry 152: 359±366 Hawton K, Fagg J, McKeown SP (1989) Alcoholism, alcohol and attempted suicide. Alcohol Alcohol 24: 3±9 Hesso R (1987) Scandinavian routines and practices in the registration of suicide. Acta Psychiatr Scand 76 [Suppl 336]: 17±21 Illsley R, Svensson P-G (eds) (1990) Health equities in Europe. Soc Sci Med [Special issue] 31: 223±430 Lereya J, Segal A, Elizur A (1995) Sudden death, neuroleptics and psychotic agitation. Progr Neuropsychopharmacol Biol Psychiatry 19: 229±241 LoÈnnqvist J, Ostamo A (1991) Suicide following the ®rst suicide attempt. A ®ve-year follow-up using survival analysis. Psychiatr Fennica 22: 171±179 LoÈnnqvist J, Tolppanen EM (1985) Outcome of attempted suicide. In: Pichot P, Berner P, Wolf R, Thau K (eds) Psychiatry ± the state of the art. Plenum Press, New York, pp 889±894 Lundberg O (1991) Childhood living conditions, health status, and social mobility: A contribution to the health selection debate. Eur Sociol Rev 7: 149±162 Mackenbach J, Kunst A, Cavelaars A, Geurts J, and the EU Working Group on Socio-economic Inequalities in Health (1997) Socioeconomic inequalities in morbidity and mortality in western Europe. Lancet 349: 1655±1659 Mehtonen OP, Aranko K, MaÈlkoÈnen L, Vapaalehto H (1991) A survey of sudden death associated with the use of antipsychotic and antidepressant drugs: 49 cases in Finland. Acta Psychiatr Scand 84: 58±64
Nordentoft M, Breum L, Munck LK, Nordestgaard AG, Hunding A, Laursen Bjaeldager PA (1993) High mortality by natural and unnatural causes: a 10 year follow up study of patients admitted to a poisoning treatment centre after suicide attempts. BMJ 306: 1637±1641 È hberg A, LoÈnnqvist J (1998) Suicides hidden among undeterO mined deaths. Acta Psychiatr Scand 98: 214±218 Ostamo A, Lahelma E, LoÈnnqvist J (2000) Transitions of employment status among suicide attempters during a severe economic recession. Soc Sci Med (accepted) Pederson AM, Tefft BM, Babigian HM (1975) Risks of mortality of suicide attempters compared with psychiatric and general populations. Suicide 5: 145±157 Petronis KR, Samuels JF, Moscicki EK, Anthony JC (1990) An epidemiologic investigation of potential risk factors for suicide attempts. Soc Psychiatry Psychiatr Epidemiol 25: 193±199 Platt S, Bille-Brahe U, Kerkhof A, Schmidtke A, Bjerke T, Crepet P, De Leo D, Haring C, LoÈnnqvist J, Michel K, Philippe A, Pommereau X, Querejeta I, Salander-Renberg E, TemesvaÂry B, Wasserman D, Sampaio Faria J (1992) Parasuicide in Europe: the WHO/EURO Multicentre Study on Parasuicide. I. Introduction and preliminary analysis for 1989. Acta Psychiatr Scand 85: 97±104 Platt S, Hawton K (2000) Suicidal behaviour and the labour market. In: Hawton K, Van Heeringen K (eds) International handbook of suicide and attempted suicide. John Wiley, Chichester, pp 309±384 Rosen DH (1976) The serious suicide attempt. Five-year follow-up study of 886 patients. JAMA 235: 2105±2109 Rudd MD, Dahm PF, Rajab MH (1993) Diagnostic comorbidity in persons with suicidal ideation and behaviour. Am J Psychiatry 150: 928±934 Rygnestad T (1997) Mortality after deliberate self-poisoning. A prospective follow-up study of 587 persons observed for 5279 person years: risk factors and causes of death. Soc Psychiatr Psychiatr Epidemiol 32: 443±450 Schmidtke A, Bille-Brahe U, De Leo D, Kerkhof A, Bjerke T, Crepet P, Haring C, Hawton K, LoÈnnqvist J, Michel K, Pommereau X, Querejeta I, Phillipe I, Salander-Renberg E, TemesvaÂry B, Wasserman D, Fricke S, Weinacker B, Sampaio-Faria JG (1996) Attempted suicide in Europe: rates, trends and sociodemographic characteristics of suicide attempters during the period 1989±1992. Results of the WHO/EURO Multicentre Study on Parasuicide. Acta Psychiatr Scand 93: 327±338 Statistics Finland (1999) Statistical Yearbook of Finland 1999. Statistics Finland, Helsinki Suokas J, LoÈnnqvist J (1991) Selection of patients who attempt suicide for psychiatric consultation. Acta Psychiatr Scand 83: 179±182 Suominen K, Henriksson M, Suokas J, IsometsaÈ E, Ostamo A, LoÈnnqvist J (1996) Mental disorders and comorbidity in attempted suicide. Acta Psychiatr Scand 94: 234±240 Suominen K, IsometsaÈ E, Henriksson M, Ostamo A, LoÈnnqvist J (1999) Treatment received by alcohol-dependent suicide attempts. Acta Psychiatr Scand 99: 214±219 Tejedor MC, Diaz A, CastilloÂn JJ, Pericay JM (1999) Attempted suicide: repetition and survival ± ®ndings of a follow up study. Acta Psychiatr Scand 100: 205±211 Valkonen T (1992) Trends in regional and socio-economic mortality differentials in Finland. Int J Health Sci 3: 157±166 Weissman MM, Klerman GL, Markowitz JS, Ouellette R (1989) Suicidal ideation and suicide attempts in panic disorder and attacks. N Engl J Med 321: 1209±1214 World Health Organization (1999) WHO/EURO Multicentre Study on Parasuicide. Facts and Figures, 2nd edn. WHO Regional Of®ce for Europe, Copenhagen
