Influence of method on seasonal distribution of attempted suicides in Italy

To verify if attempted suicides follow different seasonal distributions according to the method chosen, we studied the Italian monthly distribution of admissions for attempted suicide in the period 1984-1995 according to gender and age, distinguishing between non-violent (ICD 950-952) and violent (ICD 953-958) attempted suicides. The effect of temperature was also taken into account. A clear asymmetry in seasonal distribution of events can be observed in the oldest age groups in both genders, with a peak in warmer months. Only male attempted suicides by violent methods show a clear asymmetry in seasonal distribution, with a peak in spring months and a trough in autumn months. Spectral analysis, more accurate in detecting seasonal rhythms than analyses that only compare monthly seasonal means, identifies a circannual rhythm for violent suicides in both genders. The seasonal components of total variance in attempted suicides account for 16.5 and 12.4%, respectively, of violent male and female attempted suicides. Among males, a significant (p < 0.05) 4 cycles/year pattern is seen alongside the more relevant (89.4%, p < 0.01) 1 cycle/year distribution. Among females, aside from the 1 cycle/year distribution (63.3%, p < 0.01) noted, a 6 cycles/year pattern is found to contribute significantly (16.9%, p < 0.05) to the seasonal component of variance. A clear relationship with temperature indicators is noted among males: higher temperatures positively correlate with attempted suicides, whereas cooler temperatures seem to exert a protective action. Female attempts show a less evident correlation with indicators of temperature. Sample composition by age and gender and the lack of distinction between methods of attempted suicide may have obscured the seasonal component of some classes of attempted suicide in previous studies. The circannual rhythm found in this study of violent attempted suicide in relation to climate seasonal change confirms the importance of taking chronobiological variables into account in the evaluation of patients at risk of suicide. Copyright (C) 2000 S. Karger AG, Basel.