Seasonality in suicides: the influence of suicide method, gender and age on suicide distribution in Italy
Introduction
An asymmetrical distribution of suicide deaths in relation to the seasons has been reported for as long as reliable statistics on suicide have been available (Kevan, 1980; Chew and McCleary, 1995). In the 19th century the Italian Morselli (1881)was the first to identify this seasonal component of suicidal mortality, which he attributed to a direct influence on the circuits of the brain of climatic variables, in particular temperature. Durkheim (1897)suggested a role for factors of a relational nature; for him the spring peak in suicides observed in many Western countries depended on changes in social intercourse occurring at the beginning of the warm season.
The role of mental disorders (particularly those with a mood component) as provocative factors for suicide (Tanney, 1992) has suggested that the seasonal recurrence of disorders leading to suicidal ideation, which often come to prominence in the late winter and in the early spring months (Wehr and Rosenthal, 1989), might also contribute to the broad seasonal asymmetry in suicide deaths in both hemispheres (Goodwin and Jamison, 1990; Preti, 1997). Climatic influence on the seasonal recurrence of suicides is demonstrated by the opposite distribution patterns in the Southern and Northern hemispheres (Swinscow, 1951; Flisher et al., 1997; Yip et al., 1998). It is more difficult, however, to isolate the influence of a single climatic variable, and to separate biometeorological influence from relational influence (Souêtre et al., 1987Souêtre et al., 1990; Linkowski et al., 1992; Preti, 1997Preti, 1998).
Not all studies note a clear seasonal asymmetry of suicides. Reid et al. (1980), in a sample of Irish suicidal deaths, found no seasonal variation for either gender. Tietjen and Kripke (1994)were unable to identify a statistically significant seasonal trend for suicide for Los Angeles County, as were Zacharakis et al. (1996)for Athens. Sample composition by age and gender doubtless influences the possibility of observing seasonality in the yearly distribution of suicides; it is likely that influence of climate on suicidal behaviour is different according to the sex and age of the subject. Some studies indicate different seasonal distributions of suicides for males and females, with a single spring peak for males, and two peaks, one in spring and the other in autumn, for females (Meares et al., 1981; Näyhä, 1982, Näyhä, 1983; Lester and Frank, 1988; Micciolo et al., 1989). This autumn peak for female suicides may relate to psychosocial processes specific to Western culture, a theory based on the fact that the same trend is not observed in non-Western countries (Flisher et al., 1997; Ho et al., 1997; Yip et al., 1998).
As far as age is concerned, some studies of adolescents indicate scarce or zero seasonality in suicides, and more recent studies indicate a more evident seasonality among the elderly than among the young (Frank and Lester, 1988). Yet in New Zealand and Australia, suicidal acts among males aged 15–24 follow a single yearly cycle (Yip et al., 1998), and pronounced seasonality has been reported for American school-aged male suicides, though with a pattern opposite to that documented for older groups: among the `school-aged' group, suicides peak in winter and reach lows in summer (McCleary et al., 1991). Age-linked contingencies (in particular, the coincidence with events in the school year) have been given as explanation.
Choice of suicide method has also been found to follow a seasonal pattern. Lester and Frank (1988)observed among males spring peaks for suicide by poison, and a bimodal distribution (two peaks, one in spring and the other in autumn) for suicide by hanging or firearm. Females, instead, had spring and autumn peaks for suicide by poison and hanging, and summer and late autumn peaks for firearm suicide. In Australia Yip et al. (1998)observed a seasonal trend for suicide by hanging, but found no evidence of method seasonality in their New Zealand sample.
In Belgium, Linkowski et al. (1992)and Maes et al., 1993, Maes et al., 1994, Maes et al., 1995found that suicide by violent methods shows a clear seasonal trend, whereas suicide by non-violent methods shows no evidence of seasonality. In Israel, Schreiber et al. (1993)also observed a seasonal pattern in firearms suicide, but the yearly distribution in their sample is contrary to the distribution observed in many Western countries: colder months have a higher percentage of suicides than warmer months, and the peak for suicide is in December. Maes et al. (1995)hypothesised that the seasonal variance in suicides depends on an underlying seasonal biological variance, influencing the control of impulses. Serotonergic (5-HT) circuits are the most likely explanation for such variance, since brain levels of serotonin are sensitive to climate change, and a clear seasonal rhythm is evident in serotonergic functions, with variations which are a negative reflection of the seasonal distribution of suicide (Brewerton, 1989; Maes et al., 1995).
Low levels of serotonin are also often associated with impulsive and aggressive behaviour (Coccaro, 1989; Brown and Linnoila, 1990); it is possible, therefore, that suicide by violent methods, more often committed under the influence of strong impulsive drives, may be particularly influenced by the effects of climate on serotonergic functions.
Maes et al.'s hypothesis was based on studies concerning time series of suicides in Belgium. To ascertain if and how age, gender and method chosen for suicide also influence the seasonal distribution of events in Italy, data concerning suicides registered in Italy from 1984 to 1995 have been analyzed and their distribution by age, gender and method according to ICD 9 classification has been taken into account. The role of climate has also been considered.
Section snippets
Method
Data were taken from the records of the Istituto Nazionale Italiano di Analisi Statistiche (ISTAT) and are related to the years 1984–1995. Data for death by suicide came from judicial statistics: data were collected from the police and carabinieri (military police), who compile detailed case reports based on the preliminary death certificate completed by the examining doctor, and on extensive questioning of key informants and relevant witnesses. Classification of suicides was made according to
Demographic data
A total of 31 771 male suicides and 11 984 female suicides were identified in Italy during the 11-year study period, considering the age range of 14–65 years old and over. This corresponds to a yearly mean of 2888 male suicides (mean yearly rate, 12.6 per 100 000) and a yearly mean of 1089 female suicides (mean yearly rate, 4.4 per 100 000).
In all age groups more men committed suicide than women: 14–24 years old, ratio m:f=2625:773; 25–64 years, ratio m:f=18 427:6745; and 65 years old and
Discussion
This study shows that the seasonal asymmetry in distribution is not an attribute shared by all fatal suicidal events but varies with age, gender and suicide method chosen. For both genders a clear seasonality in the distribution of suicides, with a unimodal distribution and a peak in summer, can be found only in the older age group. Among females there is a bimodal pattern in the distribution of suicides, and an additional four-monthly rhythm is observed, both of which trends are absent among
Acknowledgements
The authors would like to thank the ISTAT staff in Cagliari for their help and kindness, Gen. Carlo Finizio, from the Meteorology Service of the Italian Air Force, who kindly supplied the climatic data, Dr. PierPaolo Pani for help in performing the Spectral analysis, and Mr. Thomas Eagle for help in the revision of the English translation. The useful suggestions by the referees are gratefully acknowledged.
References (78)
Toward a unified theory of serotonin dysregulation in eating and related disorders
Psychoneuroendocrinology
(1995)- et al.
Findings from the Seasonal Pattern Assessment Questionnaire in patients with eating disorders and control subjects: effects of diagnosis and location
Psychiatry Research
(1994) - et al.
The spring peak in suicides: a cross-national analysis
Social Science and Medicine
(1995) Physiopharmacological interactions between stress hormones and central serotonergic systems
Brain Research Reviews
(1993)- et al.
Seasonal variation of suicide in South Africa
Psychiatry Research
(1997) - et al.
Self-destructive behavior in the substance abuser
Psychiatric Clinics of North American
(1985) - et al.
Sex differences in the seasonal distribution of suicides
British Journal of Psychiatry
(1988) - et al.
Effect of some climatic factors on violent and non-violent suicides in Belgium
Journal of Affective Disorders
(1992) The influence of seasonal change on suicidal behaviour in Italy
Journal of Affective Disorders
(1997)The influence of climate on suicidal behavior in Italy
Psychiatry Research
(1998)