In Tunisia, from a legislative point of view, any violent or suspicious death poses an obstacle to the burial of the body and triggers a judicial inquiry and forensic autopsy with the aim of clarifying the cause of death and determining the circumstances of its occurrence (Official Printing Office of the Republic of Tunisia 1957). The unnatural deaths of elderly are, however, likely to be underestimated. It is because of advanced age and a chronic somatic history that clinicians will sign death certificates without the request for investigations and that autopsies will then not be practiced in several cases of suspicious deaths (Kumar and Verma 2014).
Accidents account for the majority of unnatural deaths in elderly, followed by suicide and homicide (Kumar and Verma 2014; Eilertsen et al. 2006; Akar et al. 2014). Several studies have focused on the safety of elderly people in public roads and suggest, through an analysis of risk factors, the establishment of means to protect them (Lucidi et al. 2014). In addition, injuries in the elderly proved to be more lethal than in the younger population, as even minimal lesions can be fatal (Kuhne et al. 2005) (Nagy et al. 2000). Accidents also lead to longer hospital stays, greater morbidity, complications, and readmissions (Lucidi et al. 2014) (Kuhne et al. 2005) (Subzwari et al. 2009). Elderly are people with a history of chronic diseases whose prevalence increases with age, and other dysfunctions that can impede their control of driving and can impair their ability to adapt to ever-changing urban streets (Kuhne et al. 2005). Cognitive impairments such as environmental memorization, visual disturbances, and decreased mobility and reflexes are dysfunctions that can be responsible for these dramatic ends (Eilertsen et al. 2006; Lucidi et al. 2014; Kuhne et al. 2005; Subzwari et al. 2009). Cognitive disorders were not described in our study for road accidents; however, sensory and motor impairments can be deduced from the presence of heavy and chronic antecedents. We observed that traffic accidents were more frequent on Monday but we could not determine the time of day at which they mostly occur. It has been reported, in a study conducted in Norway, that traffic accidents occur in the early morning and late afternoon due to a lower visibility (Eilertsen et al. 2006). As in our study, it was reported that pedestrians were older than drivers (Eilertsen et al. 2006). It is to note that driving while intoxicated is rare among the elderly, unlike the young population (Eilertsen et al. 2006). In our work, we have not seen any ethyl poisoning which is deleterious and plays an important role in the occurrence of traffic accidents but also drowning and falls (Eilertsen et al. 2006). This is probably due to the cultural and religious characteristics of our population, unlike other European studies. Unnatural deaths and specifically traffic accidents were higher in urban areas in this study. In contrast with these results were those obtained by Kumar and Verma in their study conducted in India (Kumar and Verma 2014).
Accidental falls are the leading cause of morbidity and mortality in the elderly (Stevens et al. 2008; John and Koelmeyer 2001). However, this varies by year and country where the studies are done. These are also the most costly accidents for health systems (World Health Organization 2007; Alamgir et al. 2012). The risk of falls increases with age (Alamgir et al. 2012; Peel 2011). The incidence would increase considerably particularly after the age of 80 (World Health Organization 2007). It is also described that falls would affect women more than men; this is explained by the difference in physical strength (Yoshida 2007) or comorbid conditions than women at the same age (World Health Organization 2007). A significant number of deaths are secondary to pulmonary embolism or pneumonia, which are complications of bed rest and immobilization due to fractures. These victims usually die in hospitals and their deaths can sometimes not be reported to the authorities leading to an underestimation of mortality due to falls (Eilertsen et al. 2006).
Burns with or without carbon monoxide poisoning accounted for 7.8% of the accidental deaths of elderly subjects in an American study (Homer et al. 2005) and 5.2% of deaths in a Norwegian study (Eilertsen et al. 2006). They are described as the major cause of accidental death in the elderly (Zhu et al. 2000). It is generally held that the elderly are prone to burns because of age-related deterioration of judgment, cognition, and mobility (Liu et al. 2013) and it has been shown that the installation of smoke detectors greatly reduces the occurrence of such accidents (Homer et al. 2005). Fire sources in our study were rudimentary heating or cooking means such as brazier used for burning charcoal, and gas cylinders feeding stoves. This is also the case for carbon monoxide poisoning that was due in our population to gas leaks from a brazier which continues to be increasingly involved in burns and accidental poisoning especially in winter. Presence of comorbid pathologies and a total area burned beyond 20% are positively correlated with death (Middelkoop and Vloemans 2015). Burns often affect the extremities including the hands, followed by the head and neck and then the trunk. They occur at home, as in cases of falls and poisoning, because elderly spend more time at home because of decrease in activity after retirement and are therefore more prone to domestic accidents. Furthermore, these accidents occur particularly in winter; this is related to the use of heating means such as charcoal or kerosene, as reported in the study conducted in China by Liu et al. (2013).
The poisoning of the elderly, unlike our study, is reported to be more often suicidal than accidental in studies carried out in India and Norway (Kumar and Verma 2014; Eilertsen et al. 2006). We observed that accidental poisoning was especially caused by carbon monoxide and due the increased use of braziers in winter. In Turkey, Akar et al. (2014) found that carbon monoxide poisoning was the second cause of death in the elderly after traffic accidents, with 21.7% and 37.7% respectively. It is noteworthy that the fatal work-related accidents of the elderly found in our study are a rare entity and very little studied in the literature. This is a rather unusual cause of death in this population, the majority of whom are retired.
The suicides of the elderly are found to be more radical, more violent, with a stronger intention (Bergman Levy et al. 2011). Non-violent means are drug poisoning, carbon monoxide poisoning, and vital submersions found more often in young people (Pitkälä et al. 2000). Hanging is the most commonly used method, followed by firearms or drug poisoning according to studies carried out in Norway and Finland (Eilertsen et al. 2006; Pitkälä et al. 2000). This is explained by the easy accessibility of hanging (Chan et al. 2007). Suicide occurs most often in the victim’s home and less frequently in public places or rest centers according to a study from Turkey (Akar et al. 2014). Male gender, retirement, social isolation, and psychiatric and somatic pathologies are predisposing factors (Akar et al. 2014; Pitkälä et al. 2000; Nomura et al. 2016). Neoplastic and neurological pathologies are particularly associated with a high risk of suicide (Pitkälä et al. 2000). Depressive disorders are more prevalent while addictive behaviors such as alcohol abuse are less common (Pitkälä et al. 2000). In addition, elderly people rarely reported suicidal thoughts to those around them. On the other hand, they present themselves to their treating physicians shortly before suicide with sleep disorder, anxiety, or depressive symptoms, which should not be trivialized. The diagnosis of a psychiatric pathology and the initiation of a treatment is indeed an effective means of prevention (Avci et al. 2017). Suicidal intoxications have not been observed in our work; the most frequently found toxic agents for some authors are antidepressants (Eilertsen et al. 2006) and for others carbamate and organophosphates (Kumar and Verma 2014).
The vulnerability of the elderly makes them also exposed to violence and homicide (Coelho et al. 2010). It is mistakenly assumed that they are respected in society and that they are protected in view of their age (Krienert and Walsh 2010). This phenomenon is increasingly encountered but still little studied (Krienert and Walsh 2010). Some American studies have shown a predominance of homicides by firearms, much less accessible in our country, with 50% of cases followed by blunt objects, then sharp tools, and finally mechanical asphyxia (Krienert and Walsh 2010; Collins and Presnell 2006). Another study from Portugal found a predominance of cranial and thoracic trauma by blunt object, as in our work, followed by sharp objects and then firearms (Coelho et al. 2010). Homicides occur most often in the victim’s home and in the context of theft (Coelho et al. 2010; Krienert and Walsh 2010; Falzon and Davis 1998). In our study, it was family conflicts that motivated half of the homicides. It is reported that men are more at risk than women, more likely to be killed by strangers, while women are more frequently assaulted by a member of the family (Coelho et al. 2010). The mean age of the victims in our study was similar to that found in the literature (Coelho et al. 2010). It is important to note that the elderly in Tunisia have recently received special protection under the law, with an increase in the sentences for violence and crimes against them, as they are considered vulnerable due to advanced age (Official Printing Office of The Republic of Tunisia 2017).
The strength of our study is that it was carried on a representative sample of the Tunisian population. It determines the unnatural manner and causes of death in the elderly and analyzes the circumstances and epidemiological characteristics of these deaths. It allows to compare our data with international literature and it permits to carry out prevention strategies in order reduce the mortality of these vulnerable individuals. The most important limitation of this study is that the number of unnatural deaths may be underestimated due to cases that can be misclassified as natural deaths and that were not referred to the department of Legal medicine. Another limitation is that our study is retrospective, data were sometimes missing, and some information could not be specified.